Factors Affecting Nesting Times in the Painted Turtle Chrysemys picta in Nebraska
Abstract
Although general descriptions of nesting in painted turtles (Chrysemys picta) are numerous, few studies have quantified the timing of the components of the nesting process, and no previous study has analyzed the factors influencing those times. The present study was conducted on the Crescent Lake National Wildlife Refuge in Garden County, Nebraska, in June 2012 and 2013 and focused on a single nesting population of female C. picta. We recorded times for the initiation of the nest foray, and the beginning and conclusion of nesting, and calculated search and construction times. Although highly variable, on average females left the marsh at 1709 hrs (n = 130), began nesting at 1745 hrs (n = 141), and finished nesting at 1915 hrs (n = 135). Search time averaged 26 min (n = 79) and nest construction time averaged 97 min (n = 104). We investigated the effects of female body size, air temperature, time of day, nest day, clutch number, substrate type, and reproductive output on these times. Smaller females nested earlier in the day, construction times were shorter in warmer temperatures, nests begun earlier in the day were completed more quickly, and nest times were longer earlier in the season and for first vs. second clutches. Soil type and measures of reproductive output had no effects on nesting times. These results suggest that temperature is a primary driver of nesting times, as these turtles exploit the afternoon–evening window of optimal temperatures, avoiding heat stress earlier in the afternoon and much colder evening temperatures as well as a presumed increase in predation risk after dark.
Population modeling studies (e.g., Crouse et al. 1987; Heppell et al. 2000; Mazaris et al. 2005; Páez et al. 2015) clearly demonstrate that the most critical life-history stage for long-term population stability in turtles is that of young adult females. For at least freshwater and marine species, the most precarious activity during this time is leaving the aquatic environment to nest (Spencer 2002; Steen et al. 2006, 2012; Refsnider et al. 2015). Hence, the components of nesting forays should be under intense selection pressure, and understanding the details of nesting in turtle populations should be a high priority.
Despite the fact that female turtles are generally at their most vulnerable while nesting, there is relatively little literature on the length of time female turtles spend on nesting forays (but see Nelson et al. 2009). Most of the research on turtle nesting forays has focused on nest-site choice (reviewed in Iverson et al. 2016). For example, the nesting biology of the painted turtle (Chrysemys picta) has been well studied for over 130 yrs (Agassiz 1857:493−500; Thoreau 1884; Babcock 1919; among many others), and many general descriptions of nesting in this species have been published (reviewed by Ernst and Lovich 2009). However, few authors have quantified the timing of the components of the nesting process (Table 1) or analyzed the factors (e.g., substrate type, female size, clutch size, ambient temperature, time of day, and weather changes; but see Christens and Bider 1987) influencing those times in painted turtles or any other species of turtle. We sought to examine nest times by quantifying nest timing and its correlates for the painted turtle at our long-term study site in western Nebraska. We hypothesized that larger, presumably older, more-experienced females would spend less time searching for and constructing a nest than would smaller, presumably younger females. We also predicted that nest construction time would be related to air temperature (faster when warmer; see Christens and Bider 1987), time in the afternoon–evening when nesting commenced (faster earlier in the day), date (slower earlier in the nesting season), clutch number (faster in subsequent clutches), substrate type (faster in sandy soil, slower in rocky soil), and reproductive output (slower in females with larger or heavier clutches or larger eggs).
METHODS
The present study was conducted on the Crescent Lake National Wildlife Refuge in Garden County, Nebraska, in May and June 2012 and 2013 and was focused on a population of female painted turtles (C. picta) in Gimlet Lake that has been under study since 1981 (Iverson and Smith 1993). Our specific study was centered in the housing area of the Refuge where nesting habitat consisted of a mosaic of nightly irrigated lawns with moderate tree cover, sandy to rocky roads, and the mowed lower margins of adjacent prairie sandhills. This population of C. picta nests only in the afternoon or evening. In addition, this population is minimally acclimated to human activities (compared with Bowen and Janzen 2008), so our observations of nesting habits were made from afar with binoculars or a spotting scope. Two to five observers monitored the primary nesting areas (∼ 12,000 m2; Iverson et al. 2016), from 1600 hrs to dark each night, from 27 May to 26 June 2012 and 16 May to 2 July 2013. Observers were dispersed across the nesting areas and in contact with the person recording time data via hand-held radio. As a result, observer effects on nesting times were minimal but not completely absent.
Time of first sight (TFS) was recorded when a female emerged from shoreline vegetation into open, mowed areas offering potential nest sites (∼ 10−20 m from the north shore of Gimlet Lake). We also recorded the nest start time (NST) when a female chose a nest site and began digging a nest pit with her back legs. Some females abandoned their first chosen site and, sometimes, subsequent sites (e.g., if she encountered a rock or was disturbed by human activity). For these females we only recorded the final NST preceding actual nest completion. Nest finish time (NFT) was recorded when a female completed covering her nest and stepped away from the site. We defined search time (ST) as NST minus TFS, nest construction time (NCT) as NFT minus NST, and total nesting time (TNT) as NFT minus NST. Nesting times for females that moved beyond the limits of our observable areas were not included in this analysis. In 2012, 67 females nested an average of 112.3 ± 60.8 m SD (range = 48.6−312.1) from the edge of the marsh and in 2013, 106 females nested an average of 88.0 ± 45.1 m SD away (range = 32.3−254.5) from the marsh.
Once a female left her completed nest, she was hand-captured, individually marked (if not previously marked), measured, and immediately released. Maximum carapace length (in mm) and spent body mass (in g) were recorded for each female. Relative clutch mass (RCM) was calculated as clutch mass divided by spent body mass; relative clutch size (RCS) as clutch size divided by spent body mass; and relative egg mass (REM) as mean clutch egg mass divided by spent body mass.
Nests were excavated immediately upon completion to determine clutch size and mass of each individual egg. Nests were then reburied and covered with hardware cloth to thwart predators. Because female painted turtles at this site frequently lay up to 3 clutches within a season, the clutch number was estimated from the nest date and the timing of previous or subsequent nests (or both) by the same female. For example, for a female found nesting only once in late May or early June, that nest was assumed to be her first clutch. For a female found nesting only once during the second half of June, we assumed she had produced 2 clutches and that we had missed her first clutch. We assumed that a female found nesting twice, in early June and late June (i.e., 24−30 d apart), had produced 3 nests, that we missed her second nest, and that the last one was her third nest (see also Iverson and Smith 1993). In 2012, we monitored 123 females (some on repeated nights) that left the marsh to nest, but only 75 of these were observed nesting. In 2013, 205 females were monitored moving to nest and only 110 were observed nesting.
Air temperature was recorded each hour at an automated National Oceanic and Atmospheric Administration (NOAA) weather station located ∼ 50 m from the approximate center of the nesting area. From those records, air temperature to the nearest 0.5°C at the midpoint of nest construction was interpolated for each nest. General soil texture at each nest site was subjectively ranked from most to least friable: 1) sandy sites had mostly open substrates with little if any grass and few to no rocks; 2) mixed sites had mostly open substrates with a mixture of sand and some buried rocks (e.g., near roads); 3) grass sites were covered by mowed lawn (often irrigated); and 4) rocky sites also had mostly open substrates often associated with roads and required females to extricate numerous small rocks in order to nest. The latter were often abandoned by females because of buried, immovable rocks (see also Mahmoud 1968).
We recorded nest time data for 76 different females over 2 yrs, including 33 recorded in both years. In addition, we recorded times for 15 females more than once in 2012 and 24 such females in 2013. To avoid pseudoreplication of data, we randomly selected only one nesting event per female within a year for our analyses (unless otherwise explicitly noted). We considered the separate years as independent sampling events.
Factors hypothesized to affect nesting times were analyzed using least-squares linear regression (for continuous dependent variables) or analysis of variance (ANOVA) (for categorical dependent variables; group distinction was determined by Fisher's Least Protected Difference). Subscripts after correlation coefficients (r) are degrees of freedom. Means are followed by ± 1 SD. Data were analyzed using Statview or Microsoft Excel.
RESULTS
The 2 yrs of study were significantly different climatically. Whereas temperatures and rainfall in May and June of 2013 were not significantly different from the 46-yr average, 2012 experienced record drought and heat (Table 2). As a result, in 2012 the first female painted turtles nested on 9 May compared with 8 June in 2013 (long-term mean = 28 May; range = 9 May–15 June; n = 23 yrs).
Body size and life-history parameters for captured turtles in 2012 and 2013 are summarized in Table 3. Only clutch size was marginally different between years (p = 0.046). Nest times for our population of painted turtles were highly variable (Table 4). Females (including repeat nesters) left the marsh on nesting forays between 1425 and 1948 hrs (mean = 1709 hrs; n = 130), began nest construction between 1451 and 2001 hrs (mean = 1745 hrs; n = 141), and completed their nests between 1613 and 2145 hrs (mean = 1915 hrs; n = 135). For the reduced data set (only one nesting event per female per year; Table 4), search time averaged 26 min (range = 1−97 min; n = 79) and nest construction, oviposition, and covering averaged another 97 min (range = 44−246 min; n = 104). None of the six nest-time parameters (TFS, NST, NCT, NFT, ST, or TNT) differed between years (Table 4). However, temperature during nesting was highly significantly different between years (Table 4), reflecting the general climate differences. Females in the warmer year 2012 began nesting forays later than in the cooler 2013.
Female body size (carapace length [CL]) was positively correlated with TFS, nest initiation, and nest completion in 2013 (r > 0.27, p < 0.04 for all cases), but not in 2012 (p > 0.21 in all cases). For the 2 yrs combined, only TFS was correlated with CL (r95 = 0.25, p = 0.015; Fig. 1). This finding suggests a trend for smaller females to begin nesting forays earlier in the day than do the larger females. Neither ST nor NCT was correlated with female body size in 2012, 2013, or in the combined years (|r| < 0.20, p > 0.17 for all cases). However, total nest time was positively correlated with body size in 2013 and for the combined years (r45 = 0.43, p = 0.002; and r74 = 0.27, p = 0.02, respectively), but not in 2012 (|r| = 0.07, p = 0.72).
Citation: Chelonian Conservation and Biology 16, 1; 10.2744/CCB-1208.1
Ambient air temperature was not correlated with TFS or NST (p > 0.13 for all cases) but was inversely related to NCT in 2013 and for the combined 2012−2013 data (r60 = −0.30, p = 0.019; and r = −0.23, p = 0.022, respectively) and nearly so in 2012 (r40 = 0.27, p = 0.08). Search time was inversely related to temperature in 2012 (r26 = −0.48, p = 0.009) but not in 2013 (p = 0.34) nor in the combined years (r75 = 0.22, p = 0.057). Construction time and TNT were each highly inversely correlated with ambient temperature (|r| > 0.42, p < 0.010 for all cases; e.g., Fig. 2). This result suggests that temperature does not influence the initiation of a nest foray, but warmer temperatures result in shorter nesting activity.
Citation: Chelonian Conservation and Biology 16, 1; 10.2744/CCB-1208.1
The TFS was not correlated with search time in 2012, 2013, or for the combined years (r < 0.16, p > 0.49). In contrast, NST was positively correlated with NCT in 2012 (r39 = 0.041, p = 0.007) and the combined years (r101 = 0.21, p = 0.037) but not in 2013 (r60 = 0.13, p = 0.30). However, in 2013 2 females were extreme outliers with construction times of 191 and 246 min. Removing those 2 records yielded a significant positive correlation with NST (r58 = 0.28, p = 0.029). Hence, nests laid earlier in the evening were constructed faster than those laid later.
Actual times for first sighting, nest start, and nest completion were not related to nest date (p > 0.11 in all cases). Search time was inversely correlated with nest date in 2012 (r27 = −0.39, p = 0.035) but not in 2013 or for 2012−2013 (|r| < 0.09, p > 0.43). Nest construction time was inversely correlated with date (Fig. 3) in 2013 (r61 = −0.34, p = 0.006; r59 = −0.78, p = 0.0003 if 2 outliers are removed) and for 2012−2013 combined (r102 = 0.28, p = 0.005), but not in 2012 (p = 0.22). Total nesting time was significantly inversely correlated with date in June in 2012, 2013, and for the combined years (r27 = −0.43, p = 0.019; r48 = −0.37, p = 0.009; and r77 = −0.39, p = 0.0004, respectively). Together, these data suggest that the time taken by females to choose a nest is probably not related to season, but the time taken to deposit a nest is generally slower earlier in the season.
Citation: Chelonian Conservation and Biology 16, 1; 10.2744/CCB-1208.1
There was no difference in actual times of first sighting, nest initiation, or nest completion for first vs. second clutches (F < 2.66 and p > 0.06 in all cases). In addition, there was no difference in search time across clutches (F < 2.56 and p = 0.12−0.82 for all cases). However, NCT was significantly longer in first vs. second clutches in 2013 (F1,61 = 6.28, p = 0.015) and for the combined years (F1,102 = 8.39, p = 0.005; 105.1 ± 38.0 min vs. 85.9 ± 28.0 min) but not in 2012 (p = 0.31). Furthermore, TNT (search and construction) was significantly longer in first vs. second clutches in 2013 (F1,48 = 6.19, p = 0.016) and in the combined years (F1,77 = 9.69, p = 0.003; 129.2 ± 39.2 min vs. 104.8 ± 29.2 min) but only approached significance in 2012 (F1,28 = 4.24, p = 0.086). These data suggest that later-season nests are not deposited at different times of day compared with early-season nests but that nest construction is faster later in the season. Soil type had no effects on nest foray timing or duration (all F < 1.5, p > 0.24).
Raw clutch size was not related to either nesting times or durations (all |r| < 0.18, p > 0.19). No relative measure of reproductive output (RCM, REM, or RCS) was correlated with search, construction, or total nest times within years or across years combined (all |r| < 0.17, p > 0.14). Except for RCM in 2013, those relative measures were also not correlated with TFS, NST, or NCT (all |r| < 0.14, p > 0.19). However, in 2013 RCM was significantly correlated with TFS (r53 = 0.28, p = 0.038), ST (r51 = 0.29, p = 0.038), and NCT (r53 = 0.31, p = 0.024). These results suggest that in a normal year, females with relatively heavier clutches may be motivated to begin their nest forays earlier in the afternoon.
DISCUSSION
Our study was complicated by the record hot and dry 2012 compared with the nearly normal 2013 (Table 2). Nevertheless, our nesting parameters were still generally similar between years (e.g., Table 4). On average, nesting at our site commenced at 1745 hrs, very similar to the weighted mean (1800 hrs) for the only other study (in Quebec) providing such data (Table 5; Christens and Bider 1987). Average time at nest completion was also similar at the 2 sites (1915 vs. 1926; Table 5), despite the differences in climate and the much larger body size of Nebraska turtles (mean CL = 177 mm) vs. those in Quebec (mean CL = 149; estimated from Christens and Bider 1986).
In contrast, we never observed painted turtles nesting in the morning at our site; however, they frequently do so in Wisconsin (Mahmoud 1968), Michigan (24%; Congdon and Gatten 1989; see also Rowe et al. 2005), Pennsylvania (30%; Ernst and Lovich 2009), and Illinois (12%; F. Janzen, pers. comm., 22 January 2016), but only rarely (1 of 53) in Quebec (Christens and Bider 1987) and apparently never in Minnesota (Legler 1954). These data suggest a longitudinal decrease in morning nesting that we hypothesize may be because of the colder early morning temperatures in the west.
Only one other study (Congdon and Gatten 1989, in Michigan) has quantified search times in C. picta and they averaged over three times longer than in Nebraska (Table 1; 102 vs. 26 min). However, nest construction times are available for several populations of C. picta (Table 1) and they are surprisingly similar at 90−100 min, despite the difference in locations (i.e., Quebec vs. Nebraska). Nest construction times are also available for a number of other turtle species (Table 6) and, not surprisingly, the species that nest in sandy soils close to water (e.g., genera Apalone, Graptemys, Malaclemys, and Pseudemys) have the shortest construction times, whereas those that nest farther from water in less friable soils (e.g., genera Emydoidea, Glyptemys, Trachemys, and presumably Mauremys), have nest construction times more similar to those of painted turtles.
Our hypothesis that larger females would spend less time searching for and constructing a nest was not supported; however, smaller females tended to initiate nesting forays (Fig. 1) and nest construction earlier in the day than did larger females. This difference could be related to inexperience of smaller, presumably younger females or to their decreased likelihood of over-heating while nesting owing to their lower surface area–to–volume ratio.
Ambient temperatures were not related to the time of first sight or nesting start time, but both search and construction times were faster and nest completion times were earlier when temperatures were warmer (Fig. 2). Similarly, as hypothesized, nests initiated earlier in the afternoon were constructed more quickly than nests initiated later. In addition, there was a weak trend for search and nest construction times to be longer earlier in the season, presumably related to the cooler ambient temperatures earlier in the nesting season, which supported our hypothesis.
We also predicted that first clutches would have longer construction times than would second clutches, which was supported in 2013 and in the combined data but not in the climatically unusual 2012. Overall, nest construction times were nearly 20 min longer in first vs. second clutches (105.1 vs. 85.9 min, respectively).
Contrary to our prediction, general soil texture had no identifiable effect on nest timing, although our results were complicated by excluding females that abandoned nest sites with buried rocks or roots. Our soil categories were also admittedly subjective but should still have been reasonable measures of friability. Presumably other variables, such as rocks or roots, clay content, soil moisture, and soil compaction, were not reflected in our soil texture categories.
Although we predicted that females with relatively larger or heavier clutches or larger eggs would take longer to nest, we found no evidence of this. However, in 2013 females with heavier clutches (relative to body mass) were more likely to begin searching for and depositing their nests earlier in the day.
In summary, our data suggest that the timing of nesting in painted turtles at our Nebraska site is primarily driven by temperature, with the initiation of nesting forays constrained by the dangers of heat stress in midafternoon and the risk of rapidly cooling temperatures in the evening, as well as probable predation risk with nightfall. The risk of predation during nesting has been shown to affect nest-site choice in some turtle species (Spencer 2002; Spencer and Thomson 2003) but not in all (Refsnider et al. 2015). In a study of relative predation pressures on female painted turtles and their nests, Refsnider et al. (2015) concluded that predation risk on nesting female painted turtles probably plays a minor role in nest-site choice and, therefore, presumably in nest timing.
It is logical that predation risk should be positively correlated with the length of time spent on nesting forays, as should darkness. However, no study has examined the risk of predation on the timing of nesting forays and nest construction. For example, it would be interesting to explore whether nest timing in turtles reflects other life-history variables (e.g., are larger species better able to exploit nocturnal nesting times). In addition, studies of the fitness consequences of nesting timing on offspring (e.g., nest survival) have also not yet been attempted. Because we protected all nests with wire screens, we could not evaluate these consequences.
Although this study oversimplified nest construction time by not parsing excavation time, egg deposition time, and nest covering time, it still quantified nesting times in much greater detail than any previous study. Quantifying these respective times in future populations will be useful, as will extending this work to other turtle populations.
Relationship of time of first sighting (in minutes after 1500 hrs) to body size (maximum carapace length in mm) for painted turtles at our study site in western Nebraska. Solid dots are from 2012 (p = 0.88); open dots from 2013 (p = 0.011). For all data combined, r96 = 0.25, p = 0.015.
Relationship of ambient air temperature (°C) during nesting to total nesting time (including search and construction times) for painted turtles at our study site in western Nebraska. Solid dots are from 2012; open dots from 2013. For all data combined, r75 = 0.54, p < 0.0001.
Relationship of nest construction time (in minutes) and nest date (day in June) for painted turtles at our study site in western Nebraska. Solid dots are from 2012 (p = 0.22); open dots from 2013 (p = 0.006). For all data combined, r102 = 0.28, p = 0.0046.
Contributor Notes
Handling Editor: Peter V. Lindeman
