Naïve Hatchlings

Eggs were collected from recently constructed nests, or by injecting gravid females with oxytocin to induce egg-laying (Ewert and Legler 1978). Eggs from each clutch were placed in damp sand inside a perforated circular plastic container (15 cm long × 10 cm in diameter) and buried in an open dune area that was fenced to exclude predators. Egg containers with partially developed eggs or hatchlings were excavated at night under a black tarp in late August and moved in light-proof containers to a dark room with the work area illuminated by a 5-W red photographer's bulb suspended directly overhead where it would provide no useful information for hatchling orientation.

Experimental Arenas

We constructed circular arenas ∼ 60–100 m in diameter, with 15-cm aluminum flashing buried 5 cm into the sand in 4 relatively flat areas in natural habitats with light and dark horizons in different directions and distances. Arena sizes were as close to 100 m in diameter as each location allowed to give hatchlings time to acquire and orient toward selected environmental cues and disperse toward a target horizon before encountering the fence. The center of the river beach arena was ∼ 35 m to the southwest of a section of the Mississippi River that provided a near open horizon and large trees to the west-southwest that provided a near dark horizon on the landward side (Fig. 1a). Two field arenas were constructed at the same location (one in spring and one in autumn) in a field that is a natural nesting area. The east and west ends of the field (and adjacent prairie to the west) represented near open horizons and to the north and south were 2 distinct near dark horizons of mixed deciduous and evergreen tree rows. No wetland or riparian habitat was visible from the arena, and the river was ∼ 275 m to the east (Fig. 2a). The third arena was located in an open area of lowland prairie with 1) a large pothole pond ∼ 75 m to the west-northwest that was not immediately visible from the arena because of a sand ridge, 2) a large pond that was visible ∼ 225 m to the north, 3) a wind row of tall pine (Pinus spp.) trees with some deciduous trees on the west end that was ∼ 230 m to the south, and 4) a large riparian forest and wetlands ∼ 280 m to the west (Fig. 3a). A fourth arena was located in an atypical upland prairie habitat with no wetland visible from the site (the closest wetland was ∼ 475 m to the west). A mixed oak (Quercus spp.)–pine forest was 60 m north of the center of the arena, and similar forests were located ∼ 130 m to the south and west (Fig. 4a). We also constructed an arena in a mature corn field with no nearby wetlands (nearest wetlands were > 950 m to the east and west); the corn stalks were ∼ 3 m high and blocked access to the visual cues typically used by naïve hatchlings (Fig. 5a).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Sixteen pitfall traps were placed level with the substrate around the interior of the fence at all cardinal directions with 0, 4, 8, and 12 = north, east, south, and west, respectively, and at the intercardinal directions with 2, 6, 10, and 14 = northeast, southeast, southwest, and northwest, respectively. Eight bucket traps were evenly spaced between the other locations. Traps were made of lengthwise halves of polyvinyl chloride tubing (30 cm long and 10 cm wide and 5 cm deep) with the ends covered with duct tape. For locating hatchlings along the fence (but not in traps), we also marked the fence at 0.2 of the distances between the lower to higher numbered traps and used those marks in recording capture locations.

Arena Protocols

Hatchlings were given an individual or release-specific paint mark and assigned to a 39-l dark gray plastic release bucket that had 4 compartments, with 2.5 × 1.5-cm escape holes covered with tape. Hatchlings from each clutch were evenly distributed among the 4 compartments and the bucket was wrapped in black vinyl material for transport to release arenas. All releases in autumn and spring were conducted at approximately the same time of day when hatchlings of all 5 species had been observed dispersing from natural nests (Congdon et al. 2011; M.J.P., unpubl. data, 2001–2003; R.M. Bolton and M. Linck, pers. comm., September 2014). At ∼ 0900 hrs on the next day with appropriate weather, one person placed the release bucket in the center of the arena, removed the black plastic covering, aligned the exit holes with the cardinal compass points, removed the tape from the exit holes, and quickly exited the area. The first direct exposure to natural light occurred when naïve hatchlings exited the buckets. Within 1 hr, we recovered any hatchlings that remained in the release bucket and removed them from the arena. We walked the fence at 30-min intervals (or more frequently during peak activity) from ∼ 45 min after release until 1 hr had elapsed without any hatchlings being captured in the evening. At each capture, we recorded the hatchling's identification, time of capture, and location (trap number or relative distance between traps). After each release was completed, hatchlings were hydrated and then released adjacent to wetlands nearest to the location of nests or where females were captured.

Statistical Analyses

Prior to analyses, all capture locations on the fence were converted to degrees by multiplying capture locations by 22.5°. Dispersal data were analyzed and Rose diagrams were created with Oriana2^TM statistical package that analyzes circular data (i.e., orientation in compass degrees; Batschelet 1981). Statistics based on circular data include calculation of mean vector of dispersal (μ), confidence intervals, Rayleigh Uniformity Test (Z), and significance level (p). When dispersal was obviously bimodal, we converted the data type from angles to axial for analyses. Rose diagrams are presented with aerial photographs (Google Earth^TM) showing the overall environmental setting of the arena, locations of wetlands, potential and environmental cues, and other features. Mean vectors of dispersal and confidence intervals are presented when directions of dispersal were significantly different from random. We use the term “directional” when dispersal patterns were significantly different from random. Measurements to wetlands and horizons were made from the center of all arenas using the ruler tool in Google Earth^TM.

River Beach (2002)

Both species of hatchling Softshell Turtles released on a nesting beach (Fig. 1a) dispersed northeast toward the open area of the Mississippi River (A. mutica, μ = 38.5°, Rayleigh-test Z = 88.2, p < 0.001, n = 97, Fig. 1b; A. spinifera, μ = 37.2°, Z = 105.0, p < 0.001, n = 132, Fig. 1c). In contrast, all 3 species of hatchling Map Turtles dispersed toward the near dark horizon of trees to the southwest and away from the river (G. geographica, μ = 202.8°, Z = 38.3, p < 0.001, n = 61, Fig. 1d; G. pseudogeographica, μ = 210.8°, Z = 59.6, p < 0.001, n = 85, Fig. 1e; G. ouachitensis, μ = 215.0°, Z = 43.2, p < 0.001, n = 58, Fig. 1f).

Field (Autumn 2002 and Spring 2006)

Hatching Map Turtles were released in the autumn when deciduous trees still had leaves and in spring before leaves developed in an arena in a mowed grain field used by nesting Map Turtles (Fig. 2a). Dispersal of all 3 species of Map Turtles was strongly bimodal and directional to the north–south toward the dark tree rows in autumn (axial: G. geographica, Z = 9.1, p < 0.001, n = 41, Fig. 2b; G. pseudogeographica, Z = 11.1, p < 0.001, n = 28, Fig. 2c; and G. ouachitensis, Z = 13.1, p < 0.001, n = 31, Fig. 2d) and spring (axial: G. geographica, Z = 23.3, p < 0.001, n = 39, Fig. 2f; G. pseudogeographica, Z = 9.5, p < 0.001, n = 15, Fig. 2g; and G. ouachitensis, Z = 3.2, p < 0.03, n = 5, Fig. 2h).

Lowland Prairie (2002)

Hatchling G. geographica dispersal in an open lowland prairie was strongly bimodal to the north-northeast toward the pond and to the south-southwest toward large deciduous trees at the west end of a wind row of pine trees (axial: Z = 4.6, p = 0.009, n = 24; Fig. 3b). The bimodal dispersal of Northern Map Turtles toward 2 far dark horizons ∼ 230 m away, but not toward the large wetland 280 m away, suggests that their perception distance was between 230 and 280 m.

Upland Prairie (2006)

Three species of hatchling Map Turtles were released in upland prairie area with no wetlands or riparian forest visible from the arena, but where some Northern Map Turtles are known to nest. Dispersal of hatchling G. pseudogeographica (Z = 9.1, p < 0.01, n = 10; Fig. 4c) and G. ouachitensis (Z = 31.2, p < 0.001, n = 41; Fig. 4d) were both strongly directional toward a nearby forest area to the north, whereas dispersal of hatchling G. geographica was bimodal, but primarily directional to the north (axial: Z = 2.7, p < 0.001, n = 19; Fig. 4b).

Corn Field (2012)

A combined sample of naïve hatchling G. pseudogeographica and G. ouachitensis were released in a mature corn field (stalks ∼ 3 m tall) that blocked access to typical environmental cues (Fig. 5a). Dispersal was primarily bimodal to the north and south and perpendicular to the crop rows, but movements in other directions resulted in overall dispersal not different from random (axial: Z = 0.94, p = 0.39, n = 41; Fig 5b).

Dispersal at a River Beach

On the same river beach used for nesting by female Softshell, Map, and Snapping Turtles, hatchling Softshell Turtles dispersed directly toward the open near horizon of the river (Fig. 1b–c) as did Snapping Turtles (Congdon et al. 2011). Approximately 2 hrs after releasing hatchlings at the river's edge, we conducted a search along the edge of the river. Hatchling Softshell Turtles were found in shallow water buried just beneath the surface of the sand and hatchling Snapping Turtles were found in clumps of submergent vegetation along the shore. Hatchling Softshell Turtles have been found in sandy stream eddies and shallows of highly dissected sandbars where their cryptic coloration and ability to bury in loose sand provides them with safe refugia (Plummer 1977). We suspect that hatchlings of both species disperse along the river edge until they find access to backwater areas with reduced current velocities and increased vegetation, similar to the habitat reported for hatchling Snapping Turtles occupying the edge of a wetland in Michigan (Congdon et al. 1993). The behaviors of hatchlings in the shallow water suggest that they are avoiding stronger currents, deep water with few resources, and perhaps the presence of large fish predators.

In contrast to Softshell and Snapping Turtles, all 3 species of hatchling Map Turtles dispersed toward the near dark horizon of trees that bordered the arena and away from the open horizon of the river (Fig. 1d–f). In addition, when all 3 species of hatchling Map Turtles from a survey of turtles in Weaver Bottoms were released in shallow water adjacent to a backwater river island, the majority left the water and moved into a nearby area of dense terrestrial vegetation (Pappas and Congdon 2003). Hatchling Map Turtles (Graptemys pearlensis and Graptemys oculifera) released at the river's edge during daylight all moved to nearby terrestrial vegetation or the landward forest 20–50 m away (Anderson 1958). Moving away from the river may result in hatchlings finding wetlands with less current and more vegetation (i.e., productivity and cover).

Autumn and Spring Dispersal of Map Turtles in a Field

Among all the release areas in this study, the field site provided hatchlings with the most distinct contrast of near dark horizons to the north and south and near open horizons to the east and west (Fig. 2a). The similar dispersal patterns of all 3 species of Map Turtles in autumn and spring indicated that seasonal differences in the tree row foliage were not sufficient to influence hatchling dispersal, and that it is the relative contrast between light and dark horizons that is the important orientation cue. In addition, a group of hatchlings observed moving from prairie areas toward backwater wetlands along the Mississippi River in spring was made up primarily of Northern Map Turtles and smaller numbers of False Map Turtles that had apparently overwintered in nests or in terrestrial refugia (M.J.P., unpubl. data, 2012–2015). In autumn releases at the same field arena, dispersal of Blanding's Turtles was also bimodal toward the dark horizons of the tree rows (Pappas et al. 2009), whereas Snapping Turtles and Painted Turtles both dispersed toward the near open horizons of the field to the east and west (Congdon et al. 2011).

Dispersal of Hatchling Northern Map Turtles in Lowland Prairie

At the lowland prairie site, hatchling dispersal was bimodal toward far dark horizons ∼ 230 m away from the arena (Fig. 3a–b). Dispersal toward the dark horizons that were closer to the arena than the large riparian forest and wetland to the west suggests a perception distance of 230–280 m, which is a distance ∼ 70 m greater than the mean distance 170 m from water for nests of G. geographica in prairie habitat. Hatchling Blanding's Turtles at the same location in 2002 initially dispersed toward the pond 225 m to the north-northeast and a large riparian forest and wetlands 280 m to the west, suggesting that the perception distance of Blanding's Turtles (325 m; Pappas et al. 2009) is greater than in Map Turtles.

Dispersal of Hatchling Map Turtles at Upland Prairie

All 3 species of Map Turtle hatchlings released at an upland prairie site dispersed directionally to the north toward the nearest dark horizon of upland forest ∼ 65 m to the north (Fig. 4b–d). Selection of a close dark horizon by hatchling Map Turtles at the upland prairie was similar to the orientation of hatchling Map Turtles at the river beach arena (Fig. 1). At the lowland prairie site, hatchling Northern Map Turtles and Blanding's Turtles (Pappas et al. 2009) dispersed toward far dark horizons.

Dispersal of Map Turtles in a Corn Field

Dispersal of a combined sample of naïve False Map and Ouachita Map Turtles in mature corn with no access to typical orientation cues was to the north and south across corn rows, which may have presented the closest approximations of near dark horizons (i.e., more sunlight would be blocked compared with gaps caused by the corn rows oriented to the east and west). Although the initial dispersals of Snapping and Blanding's Turtles in crop fields were also primarily random, some hatchlings of both species dispersed along crop rows (Pappas et al. 2013; Congdon et al. 2015). Moving in the direction of crop rows may have resulted from hatchlings following a path of least resistance (Pappas et al. 2013).

Female Northern, False Map, Blanding's, and Snapping Turtles all nest in the disturbed soils of agriculture fields before crops are mature. Mature crops block the typical cues hatchlings use for orientation when they emerge from nests, which results in many hatchlings dispersing parallel to, or away from, wetlands (Pappas et al. 2013; Congdon et al. 2015). Increased duration and distance of dispersal probably results in higher risks of injury or mortality from terrestrial predators, desiccation, or exposure to abiotic extremes (Pappas et al. 2013). Hatchling Northern and False Map Turtles that do not emerge from nests in crop fields until spring are exposed to increased risks associated with agricultural machinery during autumn harvests and soil preparation (i.e., chisel plowing) in autumn or early spring.

Distance of Nests from Water, Hatchling Dispersal Patterns, and Recruitment Habitat

The orientation and initial dispersal of hatchling Snapping and Painted Turtles were toward near open horizons regardless of whether arenas were located near or far from wetlands (Congdon et al. 2011). At one arena located close to a wetland to the west, hatchling Snapping Turtles dispersed toward open horizons to the east (away from the wetland) in the morning and to the west (toward the wetland) in the afternoon. The change in dispersal directions between morning and afternoon indicated that cues from the nearby wetland to the west were not being used for dispersal and that the difference in illumination of near open horizons associated with the position of the sun was sufficient to influence dispersal directions. We did not examine dispersal of Map or Blanding's Turtles with respect to time of day.

Hatchling Softshell and Snapping Turtles both appear to avoid strong river currents by remaining near shorelines of lotic wetlands, apparently while searching for backwater areas with reduced currents. In contrast, hatchling Map Turtles avoid lotic waters of open rivers by dispersing away from the river in search of backwater lentic wetlands with higher productivity and where they can avoid seasonal river flooding. Hatchling Blanding's Turtles that emerged from nests on shorelines of large northern lakes also dispersed inland along wetland corridors and were later found in lentic wetlands (McNeil et al. 2000).

The observation that hatchling Map Turtles on the river beach moved away from the strong currents of lotic waters suggests that they were seeking nearby lentic wetlands (e.g., small streams, swamps, marshes, beaver dams, oxbow lakes, and river backwaters) with less current, more vegetation, and higher productivity (Moll and Legler 1971; Moll 1994; Bodie and Semlitsch 2000). We examined a combined sample of all 3 species of Map Turtles during surveys conducted from 2001 to 2002 (Pappas and Congdon 2003) and additional work in 2003 (M.J.P., unpubl. data, 2003). The percentage of captures of juveniles (males and females, < 100 and 175 mm carapace length, respectively) in the river channel was significantly lower (6.0%, n = 686) compared with in river backwaters (42.2%, n = 632; G_adj = 511.3, p < 0.05).

Results from all species of hatchling freshwater river turtles support the hypothesis that they primarily use visual cues (i.e., light, dark, and contrasting horizons) while dispersing from nests. Hatchling Softshell and Snapping Turtles dispersed toward nearby open and highly illuminated areas of the river, whereas Map Turtles moved toward near dark horizons that led them away from the river in search of lentic wetlands. It is not surprising that variation in patterns of hatchling orientation and dispersal is more complex than a simple dichotomy based only on the distance of nests from water and whether hatchlings use near open or far dark horizons for orientation (Congdon et al. 2011). Hatchlings dispersing from nests located near and far from wetlands have the common goal of finding suitable recruitment habitat characterized by low current velocities and exposure to sunlight that promotes vegetation growth and increased productivity (e.g., marshes, beaver dams, oxbows, vegetated shorelines of lakes, shallow streams with low current velocities, and river areas with crab claw islands [Fremling 2005], and overwintering sites).