Climatic Warming, Sex Ratios, and Red-Eared Sliders (Trachemys scripta elegans) in Illinois
ABSTRACT
Sex ratios for hatchlings and adult turtles collected in aquatic habitats were examined for red-eared sliders (Trachemys scripta elegans) collected at Long Lake in west-central Illinois. We found that sex ratios were initially balanced but became progressively more male biased with the passage of time. Large cohorts of newly recruited males seem to underlie the increasing male bias. Recruitment more than doubled between 2001 and 2004, and these turtles were strongly male biased. Climatic warming may have led to the current male bias. A period of warming at the site has allowed females to lay more eggs by lengthening the nesting season. Females are laying an extra clutch, which accounts for the increased recruitment. This clutch is laid when soil temperatures are relatively low, explaining the male bias in newly recruited turtles. The impact of the increased number of male turtles on the population is uncertain. However, female condition declined about 7% between 1994 and 2006, suggesting that an effect may be occurring.
Sex ratio is an important aspect in the demography of all turtle species. Recognition of temperature-dependent sex determination in turtles changed the study of turtle sex ratios (e.g., Bull and Vogt 1979; Vogt and Bull 1984; Bull 1985). The discovery that the sex of hatchlings is affected by temperature has led to studies on the effects of genetic factors (e.g., Janzen 1992; Girondot et al. 1994; Morjan 2003), of maternal behavioral and physiological variables (e.g., Jeyasuria et al. 1994; Roosenburg 1996), and of environmental variables (e.g., Dalrymple et al. 1985; Janzen 1994a; Godfrey et al. 1997). Sex ratio in turtles is particularly interesting because it has a potentially direct connection to climatic variables such as global warming (Janzen 1994b).
There are many studies on turtle sex ratios and the potential factors affecting it as noted above. However, these concentrate on a particular life stage such as the hatchling or adult and may be more theoretical in their analysis. There are also many empirical studies of sex ratios in populations (e.g., Hildebrand 1932; Tinkle 1961; Dodd 1989) that investigate variation in sex ratio. Gibbons (1990) summarized the most important group of long-term studies, featuring the slider turtle (Trachemys scripta). In the current study we examined sex ratios in a population of the red-eared slider (T. scripta elegans). We report on almost 10,000 turtles collected at Long Lake in west-central Illinois. Moreover, we draw on concomitant studies of hatchlings at a nesting area near this lake and on a 13-year study of nesting females at this same lake. Thus, our study is a robust examination of the population dynamics of an important species of aquatic turtle and its potential response to changing environments. Our study is also unique in examining sex ratios during all life history stages in a single population with reference to actual climatic variables.
METHODS
Trapped Turtles
Turtles were trapped in Long Lake using hoop traps (Legler 1960) baited with fresh fish carcasses. Traps were constructed with netting with a mesh of 2.5 cm. The three steel rings of each trap have a diameter of 75 cm. Traps vary somewhat in length ranging from 100 to 140 cm long.
Long Lake is located in Jersey County, Illinois (Fig. 1) and is part of a management unit that consist of several associated lakes. The habitat including Long Lake is part of the Mississippi River State Fish and Wildlife Area and is managed by the Illinois Department of Natural Resources. Each year portions of the connected lakes exclusive of Long Lake are drawn down to promote moist-soil plant growth. During these draw-downs water levels in Long Lake tend to decline and fish kills occur. Trapping is not possible during a fish kill because turtles are uninterested in our trap baits. Thus we trapped prior to a kill (early) or after it (late) based on water depth. The depth had to be sufficient to cover the throats of the traps when the traps could be set within a meter or so from shore. Traps in this lake cannot be placed far from shore because the lake is narrow and a boating area in its center must be avoided.
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Trapping was initiated at Long Lake in 2001 and continued through 2005. Dates for trapping periods varied annually. In 2001 and 2002, our trapping period was 22 July to 10 August and 16 to 27 July, respectively. In 2005, the trapping period began 19 July and continued until 4 August. Trapping periods in these 3 years were identified as early trapping periods. In 2003, our trapping period was from 9 to 15 August, which we identified as a late trapping period. In 2004, conditions allowed trapping from 21 July to 9 September. We included an early trapping period of 21 to 30 July and then a late trapping period from 17 August to 9 September. No trapping was performed between 31 July and 16 August in 2004. Trapping dates are important because sex ratios for slider populations are known to vary seasonally (Gibbons 1990).
All trapped turtles were weighed with a spring balance to the nearest 10 g. The plastron length and carapace length, width, and height were measured with calipers to 1 mm. We used analysis of covariance to calculate least squares means for mass adjusted for plastron length for each year from 2001 to 2005 as a measure of annual turtle condition. Holes were drilled into the marginal scutes of each turtle giving them a unique code. Where possible sex was identified for all trapped turtles. Turtles identified as males had developed elongated fore claws and enlarged tail base (Ernst et al. 1994). Most males could be sexed once plastron lengths reached 100 mm (Cagle 1950; Tucker et al. 1995a). Some turtles were included that had originally been identified as juveniles but that could be accurately sexed at a subsequent recapture. Turtles with plastrons longer than 100 mm but no obvious male secondary sexual traits were considered to be juvenile females. Only turtles collected at Long Lake by trapping were included in the study. When multiple captures of an individual turtle were made in a particular year, only the first capture of the year was included in the data set. In all instances the initial capture is included even when turtles were initially identified as juveniles but were later sexed on recapture. Turtles were divided into 2 size classes for each sex at the plastron length where half the turtles were larger and half smaller (males = 141 mm; females = 170 mm).
Hatchling Study
Hatchlings were collected using a drift fence at a nesting area near Long Lake in Jersey County, Illinois (see Tucker [1997, 1999] for details). This fence varied somewhat in length but from 1997 to 2005 was 285 m long with 20 equally spaced pits to catch hatchlings. It was monitored between 1995 and 2005. Most hatchlings were released after being weighed and measured but a sample for each year from 1997 to 2005 were frozen for sex determination (Willette et al. 2005). Each hatchling was dissected and the germinal tissue visually examined under a dissecting microscope to distinguish males from females.
Nesting Females
We also report on nesting females caught from 1994 to 2006 at the same nesting area where we collected hatchlings (see Tucker et al. [1998a, 1998b]). These females were all collected as they were attempting to nest at the site. Our collecting activities at the site began in March of each year with hatchling studies. Daily visits to the nesting area began before females started to nest in May. Thus, the first nesting female for the season each year was likely to be caught. Collecting effort was kept reasonably similar in each year because once the first female was collected the site was visited daily until turtles no longer were nesting. Our interest in these turtles for this study was to estimate condition by using analysis of covariance of gravid mass adjusted by plastron length by year. Years with higher least squares means for gravid mass were years when turtles were thought to be in better condition than in years with lower least squares means. All statistics were performed using SAS for Windows (SAS Institute 2000). We calculated day of the year for any particular calendar date as the number of days elapsed between that date and January 1 with the addition of one day to account for the first day of the year for each year (range = 1–365 days; 366 for a leap year).
Climatic Conditions
Climatic data were accessed on the web site of the Illinois State Climatologist (Illinois State Water Survey 2006). The daily minimum, maximum, and average for soil temperature at 4 inches (16 cm) and at 8 inches (32 cm) were used. These environmental data were recorded at the Belleville, Illinois station about 80 km southeast of the study site. This is the station closest to the study area with a daily record for soil temperatures. We used climatic data for May through August, which is when turtle eggs incubated at our study area from 1994 to 2006. Average annual air temperatures came from the Jerseyville weather station (Illinois State Water Survey 2006), which is 16 km east of the study area. We used data collected between 1994 and 2006. Our concern was to find any recent trends in climatic conditions which could impact the turtles during the period that we studied.
RESULTS
Turtle Trapping
We gathered data on 9322 turtles collected at Long Lake (Table 1). We tested for trap bias by comparing recaptures of males and females originally marked between 2001 and 2005. Recapture rates differed significantly (χ2 = 6.30, 1 df, p = 0.0121). Females were slightly more likely to be recaptured than males (41.1% vs. 38.7%). We believe that females tend to be more trappable than males, which could lead to a slight female bias in our trapping data.
We also examined the difference in trapping dates from year to year. There was variation in trapping periods in each year (Kruskal–Wallis test, χ2 = 4195, 4 df, p < 0.0001). Sex ratios varied between early and late periods (χ2 = 9.21, 1 df, p = 0.0024). Early periods were more male biased than late periods (55.3% males vs. 52.0% males, respectively). However, this bias was not consistently present. Sex ratio for the 2004 late period was much more male biased than for the 2003 late period (53.9% male vs. 49.9% male, respectively). Thus, trapping date seemed to be a random variable. Essentially, all periods became male biased in 2004 and 2005.
The sex ratio of the turtles we trapped changed with time. From 2001 to 2005 sex ratio changed from female biased to distinctly male biased (Fig. 2; Table 1, χ2 for all). Small turtles also had a male bias, which became significant by 2003 (Table 1, χ2 for small). Large turtles became significantly male biased by year 2004 (Table 1, χ2 for large).
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Both sexes varied annually in plastron length. For males (F4,5095 = 42.3, p < 0.0001), turtles trapped in 2003 were significantly smaller than those from any other year (Table 1). Turtles trapped in 2002 were significantly larger than turtles from any other year. Turtles from 2001, 2004, and 2005 did not differ significantly from each other (Table 1). For females (F4,4275 = 59.5, p < 0.0001), turtles collected in 2005 were significantly smaller than all other years. Those collected in 2002 were larger than any of the other years (Table 1). Turtles collected in 2001 ranked second largest and were also significantly different from other years including 2002. Turtles collected in 2003 and 2004 did not differ statistically (Table 1). The associations between year and plastron length were not significant for males (rho = −0.20, p = 0.7482, N = 5) or for females (rho = −0.84, p = 0.0773, N = 5).
Recruitment seems to have increased in later years of the study. Both males and females show an increase in the number of turtles in the smaller plastron length increments (males < 142 mm; females < 171 mm; Figs. 3 and 4). The males, which start to mature and thus grow more irregularly at age 3, have a peak in small turtles with plastron lengths between 100 and 120 mm. In the early years (2001–2002) the peak was small. It was slightly more pronounced in 2003 but then increased dramatically in 2004 and 2005 (Fig. 3). Females have more regular growth into year 5. Consequently, their early small size classes show peaks for 2- to 3-year-olds and for 4- to 5-year-olds (Fig. 4). Consistent with the males, these peaks were largest in 2004 and 2005 (Fig. 4). The male bias at our site is emphasized by the difference in the scale of the y-axis between males (0–350 turtles) and females (0–160 turtles).
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
The numbers of small turtles caught was related to trap timing. For the early trapping runs in 2003 to 2005, 54% of the turtles caught were from the small size classes (females < 171 mm and males < 142 mm). For the late trapping periods from 2003 to 2005, 60% of the turtles were from the small size classes for males and females. Catching fewer small turtles would tend to underrepresent males in the early trap periods for 2005 because so many turtles were collected in that year. Early trap periods from 2001 and 2002 might also have males underestimated, but the overall impact would be small because we caught relatively few turtles in those years compared to 2004 and 2005. Regardless, the bias would be towards females, suggesting that the large male bias we found was not due to timing of trapping periods.
Comparison of mass adjusted for plastron length (an estimate of condition) varied annually for trapped turtles. Both males and females had their highest mass per plastron length least squares means in 2003 (694 g and 1043 g, respectively) that then declined into 2004 (633 g and 941 g, respectively) and 2005 (644 g and 981 g, respectively). In males, condition was not significantly associated with year (rho = −0.23, p = 0.7054, N = 5). In contrast, the decline in condition in trapped females was significantly associated with year (rho = −0.90, p = 0.0382, N = 5).
Hatchlings
Hatchling sex ratios were male biased and tended to become even more male biased in later years (Table 2, Fig. 5). The association between year and proportion of turtles that were males was positive but not quite significant (rho = 0.58, p = 0.1006, N = 9). The number of males and females identified by dissection for the years between 1997 and 2005 are listed in Table 2.
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Nesting Females
Our collecting activities at the site started before females began to nest, so the first nesting females for the season each year were likely to be caught. We found that the nesting season start date (rho = −0.90, p < 0.0001, N = 13; Fig. 6) and the nesting season length (rho = 0.066, p = 0.0184, N = 13; Fig. 7) were both significantly associated with year (Table 3). Between 1994 and 2006, the day that the first females nested became earlier and the length of the nesting season increased (Table 3). Start date was negatively but not significantly associated with average annual temperature (rho = −0.48, p = 0.099, N = 13). Nesting season length was also associated with average annual temperature and the association was nearly significant (rho = 0.57, p = 0.0521, N = 13). Thus, with increasing average annual temperature the start of the nesting season tended to be earlier and the length of the nesting season tended to increase. Comparison of the number of nesting females caught per day in 1995–2000 (Fig. 8) and to the numbers of nesting females caught per day in 2001–2006 (Fig. 9) demonstrated an added pulse of clutches early in the season.
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Although our analysis of trends in condition was not significant for trapped male turtles, it was for trapped females. However, we only had 5 years of data for trapped females. We repeated the analysis for the 13 years that we have collected nesting females. There was also a significant negative trend in mass adjusted for plastron length (rho = −0.60, p = 0.0316, N = 13 years) in nesting females. Nesting females have lost mass relative to their plastron length at a rate of about 6.6 g per year (Fig. 10).
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
Climatic Conditions
There were significant climatic trends associated with year (Table 4). Average annual air temperature at the Jerseyville weather station has increased at a rate of 0.16°C per year during the period of 1992 to 2005 (Fig. 11). Similar trends were found in the Belleville data and particularly noteworthy are changes in 4-inch and 8-inch soil temperatures in May through August, which showed significant increases during the period between 1995 and 2005. In general, air and soil temperatures have increased during the term of the study and moisture has tended to decline (Table 4).
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
DISCUSSION
We report remarkable and possibly increasingly male-biased sex ratios for the turtles that we trapped. Sex ratio is a critical demographic parameter that when sufficiently disturbed could lead to population extinction (Le Galliard et al. 2005). Gibbons (1990) and Lovich and Gibbons (1990) reviewed possible reasons for sex ratio bias in sliders. These included differential emigration or immigration of the sexes, differential mortality of the sexes, differences in age at maturity of the sexes, sampling bias, and hatchling sex ratio (reviewed by Ernst et al. [1994] as well). Our study does not consider movement patterns, but what we report is a change in demography with time. Migration and emigration patterns did not likely change during our study because the habitat at Long Lake is relatively stable from year to year.
Differential mortality may certainly be a contributor to the male bias we observed. Females, once they are mature, make long nesting journeys that expose them to many dangers (Steen et al. 2006). Males do not. Thus, some male bias is likely due to higher mortality in females. This factor would operate each year and thus cannot account for the increasing male bias. Moreover, increasing male numbers are not coming from the large turtles but instead from the smaller ones coming into the 2004 and 2005 recruitment classes. A bias in young turtles would not have any relationship to nesting because all of the young females are immature.
Age at maturity also has important effects on calculations of adult sex ratios in sliders which are often male biased (Gibbons 1990; Lovich and Gibbons 1990). However, it is vital to distinguish adult sex ratios that Gibbons (1990) was studying from a more general population sex ratio that we are studying. Adult sliders will be male biased because males mature at a younger age than do females (Gibbons 1990; Lovich and Gibbons 1990). There will therefore always be more adult males than adult females in a given population of sliders (Lovich 1996) even when the sex ratio of immature turtles is balanced. Similar biases in adult sex ratios have been reported for other turtles (e.g., Edmonds and Brooks 1996; Hailey and Willemsen 2000). Sliders at our study site differed in age at maturity, with males showing secondary sexual characteristics (elongated claws and tail base) by age three when their plastrons were about 110 mm long (Tucker et al. 1995b). The smallest gravid and thus certainly sexually mature female had a plastron length of 167 mm (Table 3; Tucker 2002). However, we did not exclude sexually immature females from our calculations. Thus, differences in age at sexual maturity would have no impact on our estimates.
Our trapping almost certainly introduced some sampling bias. Trap timing could overestimate the number of males. However, the difference would be small as comparisons of the 2004 and 2003 trapping times show. Hoop traps also can produce a male bias (Ream and Ream 1966). In contrast, we found females more likely to be recaptured than males, suggesting findings for painted turtles (Chrysemys picta) may not be applicable to sliders. Moreover, male-biased sex ratios in hoop trap collections of other species could reflect the actual population ratios (Swannack and Rose 2003). Thus in our situation, trap bias may tend to overestimate females, and the use of hoop traps is unlikely to be the reason for the male bias, we found.
Finally, we determined hatchling sex ratios for each year, and in each of the 9 years, males out-numbered females (Table 2). Since we believe that other potential sources of male bias are not important, the male bias among hatchlings is the only remaining variable left that was one of those suggested to affect sex ratios (Gibbons 1990; Lovich and Gibbons 1990). The question then becomes could the imbalance in hatchling sex ratios account for the observed sex ratio changes among trapped turtles at Long Lake?
What we suggest is that the demography in Long Lake is in flux. The numbers of small turtles that recruited into this population in 2004 and 2005 for both males and females far exceeded the numbers in 2001 and 2002 (Figs. 3 and 4). These findings cannot be due to the slight bias in catching young turtles in later trapping periods. The many young turtles that recruited in 2003 to 2005 were decidedly male biased; whereas, the smaller numbers trapped in 2001 and 2002 were less male biased (Table 1). The demographic changes at Long Lake are an increasing male bias accompanied by an apparent increase in recruitment with these new recruits even more strongly male biased.
Our findings seem to reflect 2 trends in the natural history of these turtles that appear to be responses to changing climatic conditions. One is the lengthening of the nesting season, and the other is the impact of this on the timing of egg laying. For the period that we studied local climate has warmed (Fig. 11). This may or may not be a long-term change. Between 1995 and 2006, the turtles have added as much as 21 days to the length of the nesting season by starting to nest earlier (Fig. 6). Earlier nesting may be related to the observed warming at our study area. With an internesting interval of about 2 weeks (Tucker 2001), the lengthened nesting season (Fig. 7) allows females to deposit an extra clutch each nesting season. Unlike the loggerhead sea turtle (Caretta caretta), which responded to earlier nesting by shortening the nesting season (Pike et al. 2006), the sliders in our study continue to nest into the first week of July as they did prior to the trend for earlier nesting.
The increase in recruitment that our trapping suggests has occurred would be a natural result of females laying more eggs at the site. It is unfortunate that we could not start trapping prior to 2001 because recruitment probably began to increase in 1999 when there had been a significant change in nesting dates.
The remaining question is why are so many of the new recruits males? The slider has temperature-dependent sex determination (TSD) with a pivotal temperature of about 29° to 29.5°C and is a TSD type 1a turtle with females being produced at higher temperatures and males at lower ones (Ewert and Nelson 1991; reviewed by Wibbels et al. [1998]). Increasing environmental temperatures, which we demonstrate are happening, should naturally lead to more females being produced, not more males. Janzen (1994a) found exactly that for painted turtles, another TSD type 1a species, in response to annual variation in nest temperatures. Similar females biases have been reported for sea turtles (reviewed by Blanvillain et al. 2007).
However, the timing of the clutches, we believe, is crucial to understanding how more males could be produced by a species that should produce more females with increasing incubation temperatures. We plotted actual mean soil temperatures for the nesting season and for the period immediately following it along with the pivotal temperature (Fig. 12). When nesting began in late May (1995–1997), the first clutch reached the stage critical for sex determination at a time when daily maximum soil temperatures were high enough to produce females. For the second and third clutches, there would be few days with temperatures high enough to produce many females. Given that females mostly had time to only produce 2–2.5 clutches on average (see Tucker [2001], a paper based on females prior to the full onset of the changes in nesting dates), one clutch would have produced females but the other mostly males. Likely there would have been a small male bias. One needs to keep in mind that the slider is a southern species, and it is near its northern range limit in the area we study.
Citation: Chelonian Conservation and Biology 7, 1; 10.2744/CCB-0670.1
With the lengthened nesting season (2001–2006), sliders were laying their first and third clutches at a time when there is little chance that soil temperatures, even for the daily maximums, will be high enough to produce many females in a clutch (Fig. 12). We suggest that this extra early clutch is now an extra male-producing clutch. Currently, females lay early and late clutches that should produce mostly males and only one clutch that has a good chance to produce females. Thus, the huge jump in recruitment, reflecting the extra clutch and the increase in male bias, reflecting the timing of clutch incubation relative to soil temperatures, are combining to produce the change in sex ratios that we observed. The male bias in hatchling sex ratios is consistent with the model we show in Fig. 12. Other examples of seasonal variation in sex ratios were reported for the pig-nosed turtle (Carettochelys insculpta) by Doody et al. (2004) and for map turtles (Graptemys species) by Vogt and Bull (1984).
At the present time, the degree of male bias is comparable to other samples reported by Gibbons (1990). Likely this level of bias will not have an adverse effect on the population. However, if recruitment continues at the rates we observed, conditions may become more difficult.
Turtles respond to many variables and only experimental analyses can clarify why female condition, but not that of males, seems to be declining at Long Lake. This decline, which amounted to about 7% from the highest (1995) to lowest (2004) values, coincided with the changes in climatic variables, with rising recruitment into the habitat due to increased numbers of eggs being laid, and with an increase in the number of energy consuming nesting forays. These turtles are now producing more eggs from the same resources over a longer period of time. Increased reproductive output could adversely impact female condition if energy input cannot also be increased. Consistent with our findings, male condition might be expected to show less of an effect since males do not have as large of an investment in reproduction as females. Moreover, larger numbers of males add to stress due to excessive courting, and our data for nesting females are consistent with such an effect. Le Gailliard et al. (2005) suggested that male interference with females accounted for similar mass loss in an experimental study of the effect of increasingly male sex ratios in a lizard.
Our study underscores the value of even short-term demographic studies in turtles. It also provides a note of caution when trying to predict effects of climate change on turtles with temperature dependent sex determination. The obvious expectation for a species such as Trachemys scripta would be to expect more females with increasing environmental temperatures. Instead, we suggest that the opposite has occurred because the sex is determined by a complex interaction of many variables. Future studies of turtle sex ratios would benefit from inclusions of hatchlings as well as adult turtles. Moreover, understanding nesting phenology is vital in predicting effects of climate change on turtles.
Location of Long Lake where red-eared sliders (Trachemys scripta elegans) were trapped and the nesting area where hatchlings and nesting females were collected.
Relationship between the percentage of male red-eared sliders (Trachemys scripta elegans) trapped and year at Long Lake in west-central Illinois. The rate of change is 3.29% per year, and this slope is statistically significant (R2 = 0.91, p = 0.0074).
The distribution of plastron length in male red-eared sliders (Trachemys scripta elegans) collected in west-central Illinois in 2001 to 2005. The turtles are grouped into 10-mm increments in plastron length.
The distribution of plastron length in female red-eared sliders (Trachemys scripta elegans) collected in west-central Illinois in 2001 to 2005. The turtles are grouped into 10-mm increments in plastron length.
Relationship between the proportion of hatchlings identified as male and year of collection in red-eared sliders (Trachemys scripta elegans) collected at a nesting area in west-central Illinois.
Trend in the start of the nesting season in female red-eared sliders (Trachemys scripta elegans) collected at a nesting area near Long Lake in west-central Illinois. Nesting started earlier by 2.23 days per year during the period shown, and that slope is significant (R2 = 0.80, p < 0.0001).
Trend in the length of the nesting season of red-eared sliders (Trachemys scripta elegans) between 1995 and 2006. Nesting season lengthened by 1.21 days per year and that slope is significant (R2 = 0.39, p = 0.0184).
Number of female red-eared sliders (Trachemys scripta elegans) nesting each day at the Long Lake nesting area during 1995 to 2000.
Number of female red-eared sliders (Trachemys scripta elegans) nesting each day at the Long Lake nesting area during 2001 to 2006.
Relationship between least squares means for mass adjusted for plastron length and year in nesting female red-eared sliders (Trachemys scripta elegans) collected at a nesting area near Long Lake in west-central Illinois. The turtles are losing an average of 6.6 g per year, and that slope is significant (R2 = 0.30, p = 0.0313).
Recent trends in average annual temperatures at the monitoring station in Jerseyville, Jersey County, Illinois, for the years between 1992 and 2005. The regression is significant (slope = 0.12°C per year, R2 = 0.39, p = 0.01).
A model showing actual means by week from May 1 for 4-inch soil temperatures measured at Belleville, Illinois, with weeks that clutches are laid, assuming a minimum 2-week period between sequential clutches for female red-eared sliders (Trachemys scripta elegans) laying 3 clutches and the time period during incubation that sex determination would be likely to occur.
