Methods

Our study was conducted in 2002 along a 3.8-km stretch of stream that terminates at the Delaware River in Warren County, New Jersey. The stream was bordered by deciduous forest composed primarily of silver maple (Acer saccharinum), black walnut (Juglans nigra), oaks (Quercus spp.), and agricultural fields. The agricultural fields were used to grow corn (Zea mays) and included 42% (66 ha) of the total site area. The farming schedule practiced here was soil tillage in April, planting in May, fertilizing and herbicide application in June, and harvesting in October. Turtles used the cornfields for nesting during the last week of May and first 2 weeks of June. The majority of hatchlings emerged in mid to late August.

We installed nest protectors constructed of 2.5-cm wire fencing over 4 nests at the time of egg deposition (Graham 1997). The mean straight-line distance between the nests and the nearest aquatic habitat was 67 ± 21 m (mean ± 1 standard deviation; range 49–90 m). We monitored the nests throughout the incubation period and later weighed and measured hatchlings within 24 hours of their exiting the nest chamber during emergence. Ten hatchlings from the 4 nests were fitted with BD-2A transmitters (Holohil Systems Ltd, Carp, Ontario, Canada) with an expected battery life of 2–3 weeks and signal range of approximately 500 m (Fig. 1). Transmitters were glued to the carapace using Devcon 5-min epoxy (ITW Brands, Wood Dale, IL) and removed or replaced after 15–19 days, near the anticipated end of battery life. The total package mass of the transmitter and epoxy averaged 0.8 g and was 8–10% of neonate weight. We did not have enough transmitters to replace all of them on all hatchlings until they entered water; on 3 hatchlings the transmitters were replaced once, and on one it was replaced twice. These 4 individuals were weighed and measured when the original transmitters were removed and before the replacement transmitters were attached. The site was visited between 0700 and 1900 hours (EST) from 1 August to 15 October 2002. The mean time interval between tracking locations was 1.5 ± 0.1 day (range 1–9). The position change of hatchlings relative to their last location was determined by Universal Transmercator coordinates at each capture location using a Garmin III Plus handheld Global Positioning System unit with an error of approximately 3 m (Garmin International Inc., Olathe, KS).

View Figure

• Download as Powerpoint
• Download Figure

To compare behavior between age classes, we fitted 10 adult G. insculpta (5 males and 5 females) with R1535 transmitters (Advanced Telemetry Systems, Isanti, MN) and tracked them simultaneously with hatchlings. We recorded turtle behavior at each capture location as resting (beneath vegetation or in forms), basking (fully or partially exposed), foraging (prey on beak or in mouth), or traveling (moving between locations). We measured air temperature at 1 m above ground for each capture location using a rapid register thermometer (Miller & Weber Inc, Ridgewood, NY). Substrate temperature was measured at the surface of the ground beside each individual. We used Student t-test to determine if air and substrate temperatures differed between hatchling and adult G. insculpta locations; significance was accepted when p < 0.05. We report the results as means ± 1 standard deviation.

Results

Hatchlings emerged from the nests from 13 to 20 August after incubation periods of 69–76 days (Table 1). The average mass and straight-line carapace length of the 10 hatchlings prior to being fitted with transmitters were 8.9 ± 0.6 g (range 8.0–10.1 g) and 37.9 ± 1.1 mm (range 36.0–39.2 mm), respectively. Three hatchlings were lost within 24 hours. The mean tracking period for the remaining 7 was 30 ± 17 days (range 13–62 days), and the mean total straight-line distance traveled was 493 ± 306 m (range 163-1052 m) (Table 1). The average minimum daily movement (straight-line distance between location sites) was 24 ± 8 m (range 7–34 m), expressed as the mean of the means for each hatchling for observations separated by 1 day; if mean daily movement is calculated as (total straight-line distance moved)/(total days tracked), this figure is 18 m/d. Movements were greatest during the first 24 hours, when they averaged 131 ± 84 m (range 2–257 m), which could be an artifact due to the handling associated with transmitter attachment. Movements varied greatly among and within individuals. For example, turtle 2 made an initial movement of 81 m and then stayed in the same area for 14 days. Four hatchlings remained in the agricultural fields during the tracking period, while 3 reached water (Delaware River or parental home stream). The mean travel time and total distance traveled by these latter 3 were 43 ± 17 days (range 29–62) and 546 ± 457 m (range 163–1052). This distance was 8 times greater than the mean straight-line distance between the nests from which they emerged and the nearest water.

Table 1. Movement data for 10 G. insculpta hatchlings following emergence from nests in New Jersey.

View Fullscreen

When located during the day, the hatchlings were found resting 57.6% of the time, basking 36.8%, foraging 4.0%, and traveling 1.6%. Resting hatchlings were located primarily beneath dead corn leaves and stalks (from the previous season) but also beneath flattened patches of green corn, yellow wood sorrel (Oxalis europaea), woodland horsetail (Equisetum sylvaticum), dewberry (Rubus flagellaris), carpetweed (Mollugo verticillata), yellow nutsedge (Cyperus esculentus), black nightshade (Solanum nigrum), and poison ivy (Toxicodendron radicans). Basking hatchlings were found partially covered by vegetation or adjacent to the prop roots of standing corn. Traveling hatchlings were observed walking between cornrows.

Hatchlings were observed foraging 7 times on slugs (Arion subfuscus); 6 of these events occurred on overcast days with light to heavy rain. After 12 to 18 days from release, 4 hatchlings increased their carapace length an average of 2.6 ± 0.8 mm (range 1.7–3.4 mm) and mass 1.1 ± 0.3 g (range 0.9–1.5 g). Hatchlings that reached water were found submerged beneath branches, duckweed (Lemna sp.), and floating mats of Elodea, at depths of 0.1–0.5 m. Average water temperature for these observations was 22.3° ± 2.8°C (range 19.5°–25.0°C).

The average mass and straight-line carapace length for the 10 adults radiotracked simultaneously with the hatchlings were 892.9 ± 149.3 g (range 735.0–1190.0 g) and 192.9 ± 14.5 mm (range 168.8–220.0 g), respectively. The total number of tracking locations for adults was 202. The percentages of observations made on the diurnal activity of adults were 49.1% resting, 22.7% basking, 21.1% traveling, and 7.1% foraging. Air temperatures at the capture sites for adults were significantly (t = 4.01, df = 15, p = 0.001) higher (28.1° ± 1.7°C, range 24.7–31.5°C) than those for hatchlings (25.3° ± 0.8°C, range 23.9°–26.3°C). Likewise, substrate temperatures at the capture sites for adults were significantly (t = 5.62, df = 15, p < 0.001) higher (26.4° ± 1.1°C; range = 24.4°–28.0°C) than those for hatchlings (23.5° ± 1.0°C; range = 22.5°–25.1°C).

Discussion

Some species of turtles that typically live in or near water can feed on land as well as in the water, including G. insculpta, E. blandingii, and G. muhlenbergii. For these species, moving directly to water after emergence from the nest may not be as crucial as it is for species that can only feed in water (e.g., Apalone mutica and A. spinifera; Doody 1995). Perhaps for hatchlings of the more terrestrial species, it is only necessary to eventually move to water for hibernation. Wood turtle hatchlings in New Jersey spent 4–12 weeks on land where they fed and grew. Our findings are consistent with those of Tuttle and Carroll (1997, 2005). They found that G. insculpta hatchlings in New Hampshire could take as long as 26 days to enter a brook, and would sometimes move away from the nearest brook rather than toward it. They also found that of 12 turtles that were tracked using fluorescent powder from their nest to water, transit times averaged 6.5 days (range 1–24 days). Similarly, E. blandingii hatchlings in Nova Scotia have also been found to tarry on their nest to water journey and even to actively avoid large bodies of water adjacent to their nests in favor of more distant, shallower waters (Standing et al. 1997; McNeil et al. 2000). In Massachusetts, E. blandingii hatchlings moved more directly to water, but the nests were near wetlands and transit time was as long as 9 days (mean 2.9 days; Butler and Graham 1995).

For the few aquatic species that can and do feed on land, there is no immediate need to enter water for feeding; although, desiccation and predation are still potential threats. At least 7 of the 10 neonate G. insculpta that we radiotracked remained primarily terrestrial following emergence, and they fed and grew during the period they spent on land. While an increase in mass might be attributed to water gain, the increase in carapace length indicates that growth was occurring prior to hibernation. Our study and those of Tuttle and Carroll (1997, 2005) found that hatchlings spent the night in forms. All 3 studies found most hatchling activity to occur in the morning, and activity to be stimulated by rainfall. These observations are consistent with avoidance of potential desiccation associated with higher midday temperatures, as is our finding that adult turtles are more active at higher temperatures, which would not pose as great a threat of desiccation to them because of their larger size. Ernst (1968) noted that young turtles, including G. insculpta, experienced greater water loss than older individuals in laboratory experiments and suggested that this difference was due to the greater surface to mass ratio of smaller individuals. In addition, Tuttle and Carroll (2005) reported that at least one hatchling made a “beeline” for a brook, indicating that remaining on land for protracted periods after emergence is not an inviolable rule. Three of the 10 hatchlings that we released were not located again, and it is possible that they were either taken by predators, were lost due to transmitter failure, and/or they moved directly to water and then moved far enough downstream to be out of receiver range.

The average time spent on land by 12 G. insculpta hatchlings in New Hampshire that were followed to water was 6.5 days, with a maximum of 24 days; another was found still on land after 26 days (Tuttle and Carroll 1997, 2005). At our New Jersey site, the mean time spent on land for the 3 hatchlings followed to water was substantially longer (mean = 43 days, maximum = 62 days). Since the climate is more moderate in New Jersey than New Hampshire, we assume that the ultimate factor determining movement to water is the necessity of aquatic hibernation, and the cue to leave the terrestrial environment is falling temperature. It has been demonstrated that extrinsic loads may impair locomotion in turtles (Zani and Claussen 1995). Thus, transmitters could have affected the travel rate of turtles in our study; however, the ratio of transmitter weight to body weight did not exceed the recommended maximum for reptiles (10%; Anonymous 1987). Furthermore, the average daily movement of turtles in New Jersey (24 m) was the same for turtles in New Hampshire (Tuttle and Carroll 2005).

While it is well documented that hatchlings of several species of turtles overwinter terrestrially in or below the nest cavity without emerging above ground, there are no data to suggest that hatchling turtles in cold-temperate climates that do not overwinter in the nest routinely overwinter on land. However, terrestrial overwintering may be possible under certain conditions for a few individuals that do not reach aquatic hibernacula before the advent of cold weather (Congdon et al. 1983, 2000; Standing et al. 1997, 1999; McNeil et al. 2000; Pappas et al. 2000; Dinkelacker et al. 2004). Now that radiotransmitters are manufactured small enough to place on neonatal turtles, it is possible to study the behavior of hatchlings between the time of emergence from the nest and their first hibernation period. Moreover, researchers now have the capability to identify the locations and characteristics of aquatic hibernacula, which are currently undescribed but likely to be different from those of adults (Reese et al. 2004; Ultsch 2006).

The terrestrial behavior of hatchling G. insculpta following emergence from nests in agricultural fields has implications for the conservation and management of this species. The prolonged stay of hatchlings in this habitat following emergence might put them at risk of injury or death inflicted by farm machinery if crop harvest occurs before they have migrated to aquatic hibernacula. Injuries and deaths caused by farm machinery have been reported for adult G. insculpta at this location (Castellano 2007) and elsewhere (Saumure and Bider 1998; Saumure et al. 2007). At our New Jersey site, late-maturing varieties of corn are grown in order to delay harvest until mid-October when both adult and neonate turtles have exited the fields. A management scheme to delay harvesting until turtles have entered aquatic habitat for hibernation is advised for other locations.