Experimental Release Arenas

Seven circular arenas approximately 90 m in diameter were constructed with 15-cm aluminum flashing buried 5 cm into the sand at 6 different locations on relatively flat terrain in the field. Orientation patterns of naïve hatchlings were examined in arenas constructed in 1 natural nesting location during spring (painted turtles) and in 4 natural nesting locations and 2 atypical locations in fall (snapping turtles; Table 1). Locations on the arena perimeter fences that correspond to directions from the center of the arena were marked with evenly distributed numbers from 0–15. Numbers 0, 4, 8, and 12 were located at the cardinal directions, north, east, south, and west, respectively; 2, 6, 10, and 14 at the primary intercardinal directions, 45°, 135°, 225°, 315°, respectively; and 1, 3, 5, 7, 9, 11, and 13 were evenly spaced among the primary intercardinal directions.

Table 1 Hatchling release sites and locations.^a

View Fullscreen

We use the terminology used in other studies of hatchling freshwater and sea turtle orientation to describe the areas that surround the arenas (Anderson 1958; Mrosovsky and Carr 1967; Tuttle and Carroll 2005; Pappas et al. 2009) where “horizons” refer to “the panoramic view or angular degrees of the landscape…” (Olden et al. 2004). Horizons and horizon contrasts are specific to each location, with fields and prairie habitats categorized as illuminated, open horizons, and contrasting riparian and upland forests as dark horizons. Near and far horizons were based on their relative distances from each arena (i.e., a far horizon at one arena may be closer than one at another location).

Arenas were located on relatively level terrain to exclude geotaxis as a major factor and to allow hatchlings immediate access to all near- and far-horizon cues (Table 1). We present information regarding situations, goals, and expectations for specific arenas at the beginning of each section of the results.

Hatchlings

Eggs were obtained directly from gravid female snapping and painted turtles by using oxytocin induction (Ewert and Legler 1978) or from natural nests. The eggs were incubated in perforated plastic containers (15-cm height by 10-cm diameter) labeled with the date of collection, female identification or nest number, and clutch size. The containers were buried approximately 10 cm deep in an area of nesting dunes that was protected with hardware cloth. Eggs or hatchlings were excavated at night under a tarp in late August for snapping turtles (which typically emerge from nests in fall) and in spring for painted turtles (which typically emerge from nests in spring). All the hatchlings were transported to a dark room in light-proof containers.

Hatchlings were given nest cohort paint marks in a dark room illuminated by a 5-W red photographer's bulb located directly overhead. On the evening before release, clutches of hatchlings were distributed as uniformly as possible among 4 equally sized compartments within each release bucket. Each compartment had a 3-cm diameter escape hole covered with tape to prevent hatchlings from exiting until they were at the release area. On the day of release, the buckets were completely double wrapped in black vinyl before transport to the release arenas.

Release Protocols

The majority of releases of hatchlings were conducted in the morning, and, when sample sizes allowed, we conducted subsequent morning releases on the following day. In conjunction with morning releases at 3 arenas, the hatchlings also were released in late afternoon. Notations (R1, R2, or R3) each indicate a set of hatchlings of a given species simultaneously released at the same arena. With the exception of 2 releases conducted on mostly overcast days at McCarthy in 2002 (R3) and 2006 (R1), all other releases of snapping turtles were conducted during sunny days in early September when hatchlings were dispersing from natural nests.

Between 0830 hours and 0900 hours (morning releases) or 1530 hours and 1630 hours (afternoon releases), one person placed a release bucket in the center of an arena, removed the black plastic covering, and aligned the 4 exit holes to the cardinal directions to evenly distribute the initial direction of hatchling emergence. Tape was then removed from the exit holes, and the person quickly left the area. After 1 hour, the bucket was checked to ensure that all the hatchlings had exited (the few hatchlings that failed to exit the release buckets were removed and released in the closest wetland to the location of their nests or where the female parent was captured). The first exposure to environmental cues occurred when naïve hatchlings exited the buckets.

We walked the fences at 15–20-minute intervals, starting approximately 30 minutes after hatchling release and continuing into the evening until an hour had elapsed without any hatchling activity. At capture, hatchling identification, time, and fence number (estimated to the nearest 0.1 of fence numbers) were recorded. We terminated experiments after the majority of the hatchlings were captured at the arena fence or at the end of the second day (except for morning and afternoon releases at Mallard 2008 arena, where hatchlings were recaptured over 3 days). All hatchlings used in the study were released in wetlands near the location of their nest site, or their female parent's point of capture.

Statistical Analyses

Before analyses, all capture locations on the fence were converted to degrees by multiplying capture locations by 22.5°. Mean vectors of dispersal and 95% confidence intervals are presented for unimodal dispersals only by using the Oriana2 program for circular statistics (sample sizes are presented in Table 1). Proportions of hatchlings dispersing in different directions were made by using adjusted G tests. Dispersal vectors are presented as Rose diagrams (Oriana2) with aerial photographs of release areas and locations of arenas. Straight-line distances from arenas to features of the release sites were measured with the ruler tool in Google Earth.

West Newton Field

West Newton Field (Fig. 2a; morning releases of painted turtles in spring 2006 and snapping turtles in fall 2002) presented hatchlings with clear choices between 2 nearby open horizons, east toward the Mississippi River (275 m and not visible from the arena) and west toward nearby elevated dunes (125 m and away from any wetlands). Two equidistant dark near horizons 60 m to the north and south from the center of the arena were composed of thick tree rows of mixed pine and deciduous trees. Based on the results of Noble and Breslau (1938), our expectations were that 1) hatchling snapping and painted turtles would primarily move toward the open areas to the east (toward the river) and/or west (toward the less open and elevated dunes).

View Figure

• Download as Powerpoint
• Download Figure

Orientation of hatchling painted turtles in R1 and R2 in spring 2006 were similar and were significantly different from random (Rayleigh-test, Zs > 16.8, p < 0.001; Fig. 2b, c). Mean vectors were to the east (R1  =  101.3°, R2  =  99.1°). For R1 and R2 combined, the proportions of painted turtle hatchlings dispersing toward the open areas to the east and west were 78% and 3%, respectively, and dispersing toward the dark tree rows to the north and south were 7% and 12%, respectively.

The dispersal of snapping turtle hatchlings in releases R1 and R2 in 2002 both had primary vectors to the east and much smaller secondary vectors to the west, and both releases were significantly different from random (Rayleigh-test, Zs < 3.40, p < 0.0002; Fig. 2d, e). For R1 and R2 combined, the proportions of snapping turtle hatchlings that oriented toward open areas to the east and west were 58% and 35%, respectively, and those dispersing toward the dark tree rows to the north and south were 0% and 7%, respectively.

Overall, the majority of painted and snapping turtles dispersed toward open near horizons to the east and west. A primary difference between the dispersal patterns of the 2 species was that a larger proportion of snapping (35%) vs. painted turtles (3%) moved west toward the dunes (G_adj > 33.2, p < 0.05), and secondarily a larger proportion of painted turtles (19%) vs. snapping turtles (7%) dispersed toward the tree lines to the north and south (G_adj > 6.5, p < 0.05).

West Newton Beach

The West Newton Chute of the Mississippi River created an open and illuminated near horizon between north and east from the center of the arena, and a dark near horizon of the onshore woods extended in almost all other directions. Both vision and olfactory cues were potentially available to snapping turtle hatchlings because the river was only 60 m from the center of the arena. Our expectation was that hatchlings would move directly toward the nearby open area of the beach and river.

Orientation of 29 naïve hatchlings released in the morning was significantly different from random (Rayleigh-test, Z  =  15.74, p < 0.001) with a mean vector of dispersal to the northeast (51°, confidence interval ± 35.1°) toward the open area and river (Fig. 3; fall 2002). In spite of the proximity of the river, the initial dispersal of 14% of the hatchlings was parallel to, or away from, water.

View Figure

• Download as Powerpoint
• Download Figure

The Nature Conservancy Prairie

Both arenas were in relatively flat sections of the prairie used for nesting by snapping turtles, and adults and hatchlings occupy the closest wetland, a pond bordered with a thick row of alders 240 m and 110 m west of the arenas in 2006 and 2008, respectively (Fig. 4a; morning release in fall 2006 and morning and afternoon releases in 2008). A more distant, extensive wetland bordered by a deciduous riparian forest was 330 m to the south southwest of both arenas. A stand of large pine trees 95 m north of the 2008 arena represented a dominant dark horizon. In 2006 and 2008, we expected that dispersal of hatchlings would be less directional than at West Newton Field or West Newton Beach sites, because nearby horizons were open in all directions (Fig. 4a). Because the 2008 arena was located with the dark band of alders around the pond to the west and a dark nearby horizon of pines to the north, we expected orientation to be primarily toward the open areas to the south or southeast. In 2008, we also released hatchlings in the morning and late afternoon to determine if the time of day would influence orientation.

View Figure

• Download as Powerpoint
• Download Figure

In 2006, initial heading of hatchlings was nonrandom (Rayleigh-test, Z  =  17.9, p < 0.001) and to the southwest toward a nearby large open area (mean vector, 235°) and in line with a wetland 330 m away (Fig. 4a, b). The proportions of hatchlings that dispersed away from wetlands (0°–180°) was 11% (Fig. 4a, b).

In 2008, the initial dispersal of hatchlings in the morning and afternoon (Fig. 4c, d) were both nonrandom (Rayleigh-test, Zs > 6.61, p < 0.001). The majority of hatchlings released in the morning primarily moved to the south (mean vector, 179°), whereas dispersal of those released in the afternoon was less directed (mean vector, 296°) to the northwest.

Department of Natural Resources Prairie

The arena (Fig. 4; fall 2007 morning and afternoon releases) was located in a natural nesting area on relatively flat terrain with a wetland < 100 m away to the west and a stand of large pine trees 75 m to the south. Hatchlings released in the morning and afternoon had exposure to open horizons to the north, east (rising sun and away from wetland), and west (setting sun and toward the wetland). Because of the proximity of the wetland, we expected that hatchlings would move west toward the open horizon and wetland, regardless of time of day.

Orientation of hatchlings released in the morning and afternoon were both nonrandom (Rayleigh-test, Zs > 5.0, p < 0.006; Fig. 4e, f). Orientation of hatchlings released in the morning was bimodal with a majority of hatchlings moving east and away from the wetland and primarily to the west toward the wetland in the afternoon. The proportions of hatchlings moving away from the wetland in the morning (65%) and afternoon (34%) were significantly different (G_adj  =  15.4, p < 0.05).

Mallard

Compared with all other release areas, the Mallard arena (Fig. 5a; fall 2008 morning and afternoon releases in an atypical area) was the most distant (475 m) from any wetland. We expected that the orientation of snapping turtles 1) would be random because the arena was surrounded by nearby open horizons, and 2) based on results from earlier releases, would be influenced by the time of day when hatchlings were released.

View Figure

• Download as Powerpoint
• Download Figure

Orientation of hatchlings released at Mallard in the morning (Rayleigh-test, Z  =  24.33, p < 0.001) and afternoon (Rayleigh-test, Z  =  5.63, p < 0.003) were both nonrandom. In contrast to The Nature Conservancy (TNC) and Department of Natural Resources (DNR) Prairie sites where time of day strongly influenced orientation; mean vectors of dispersal for hatchlings released in the morning and evening were both south at 168° and 198°, respectively (Fig. 5b). Approximately equal numbers of hatchlings released in the morning were recaptured over each of 3 days while walking the fence, whereas 68% of those released in the afternoon were recaptured over 2.5 hours (1745–1915 hours) on the day of release.

McCarthy

The arenas at McCarthy (Fig. 6a; 2002 and 2006 morning releases in an atypical area) were more distant (275–350 m) from wetlands than are typical snapping turtle nesting areas at Weaver Dunes. One major feature of the horizon at McCarthy was a windrow of large introduced pine trees that were 200 m and 90 m south of the arenas in 2002 and 2006, respectively. We expected that 1) hatchlings in all releases would not respond to visual cues from the distant McCarthy wetlands, 2) hatchling dispersal in 2002 would be random, because nearby open areas were located in all directions, and 3) dispersal in 2006 would be any direction but to the south because of the proximity of the dark horizon of the tree row.

View Figure

• Download as Powerpoint
• Download Figure

In 2002, R1 and R2 were conducted on sunny days, and orientations of hatchlings were both significantly different from random (Rayleigh-test, Zs > 9.5, p < 0.001; Fig. 6b, c) and bimodal to the southeast and southwest toward open prairie. Orientation of hatchlings released on a mostly overcast day (R3) also was bimodal and nonrandom (Rayleigh-test, Z  =  4.6, p < 0.01; Fig. 6d), but primarily to the east and secondarily to the west, and a smaller proportion of hatchlings dispersed south toward the pine trees than in releases R1 and R2.

In 2006, orientation of hatchlings was nonrandom (Rayleigh-test, Z  =  8.1, p < 0.001), with the majority of hatchlings dispersing in a westerly direction (mean vector, 238°; Fig. 6e). Five hatchlings (17%) moved south toward the dark tree row.

First-order Estimate of Perception Distance of Snapping Turtle Hatchlings

Fewer than 4% of the hatchlings in R1 and R2 combined at West Newton Field (Fig. 2a) moved toward the tree rows 65 m to the north and south, and, at the DNR Prairie (Fig. 4a), only 2% moved toward the pine stand 75 m to the south. The small number of hatchlings that moved toward the trees at both sites suggests that the dark horizons were visible to the hatchlings and were being avoided. In contrast, during the afternoon release at TNC (Fig. 4a, d) in 2008, a large group of snapping turtle hatchlings moved directly toward a dark horizon created by a stand of pine trees 95 m north of the arena. Even though the 2006 arena at McCarthy (Fig. 6a) was approximately half the distance to the trees (90 m) compared with the 2002 arena (210 m), the proportions of hatchlings that dispersed toward the dark tree row to the south of the arena were similar. The results from both years are most consistent with the hatchlings orienting toward open horizons rather than avoiding the tree row (i.e., the tree row 90 m away was beyond the hatchlings perception distance because it was at the TNC site). Because orientation at all other locations and releases was primarily toward nearby illuminated horizons, hatchlings probably dispersed toward the open area between the arena and the tree stand to the north rather than perceive and avoid the dark horizon created by the trees (Fig. 4a–d). Collectively, the orientation of hatchlings at 3 arenas suggests that the perception distance of hatchling snapping turtles was > 50 m and < 90 m.

Hatchling Orientation at Release Sites

At West Newton Field (Fig. 2), the major direction of dispersal of painted turtles was east into an open area and toward the river (not visible from the arena). The majority of snapping turtles also dispersed to the east, with a secondary direction to the west. Only a few individuals of either species dispersed toward the dark tree rows. That the 2 equidistant dark near horizons composed of thick tree rows 60 m to the north and south of the center of the arena were avoided by hatchlings suggests that they were within the perception distances of both species.

At West Newton Beach (Fig. 3), the majority of snapping turtles dispersed toward the open horizon created by the river; however, because of the close proximity of the water, we could not rule out olfaction or a humidity gradient as an influence on the orientation of the hatchlings.

Patterns of orientation of hatchling snapping turtles released in the morning at the TNC Prairie in 2006 and 2008 (R1) were all toward nearby open areas but not toward the closest wetland (Fig. 4a–c). In 2008, differences in illumination caused by the position of the sun in the morning and afternoon apparently altered the hatchlings perception of openness and intensity of illumination of different patches of prairie. A large group of hatchlings released in the afternoon in 2008 appeared to atypically disperse toward a dark far horizon of pine trees 95 m north of the arena (Fig. 4a–d). However, orientation patterns at all other arenas suggest that hatchlings were actually dispersing toward the near horizon of open prairie between the arena and the trees (i.e., the stand of trees were outside of their perception distance).

The morning and afternoon releases of hatchling snapping turtles at the DNR Prairie site in 2007 (Fig. 4) resulted in the most dramatic change in orientation associated with time of day (see section on polarized light below). Patterns of orientation and dispersal of hatchlings released in the morning and afternoon were essentially mirror images of each other (Fig. 4e, f). Dispersal was primarily to the east (toward the rising sun and away from the wetland) in the morning and to the west (toward the setting sun and the wetland) in the afternoon; however, both directions were toward nearby open patches of prairie. The direction of hatchling dispersal in the morning was not confounded with reflected light as a dispersal cue because the arena was between the sun and the wetland. Orientation in the afternoon release was the only result that could be considered consistent with the detection of reflected polarized light because the wetland was between the sun in the west and the arena to the east.

In contrast to releases at the TNC and DNR sites, hatchling orientation and initial dispersal in morning and afternoon releases were toward the same open area to the south at Mallard (Fig. 5). There were no obvious features (e.g., vegetation or topography) that appeared to account for the lack of response to time of day at the Mallard Dune site compared with the substantial differences in orientation that occurred at the TNC and DNR Prairie sites.

Proportion of Hatchlings That Did Not Move Toward Open Areas

In a variety of settings, the proportion of hatchlings that did not move toward open areas ranged from 8% to 17%. If orientation of hatchlings at West Newton Beach had been reinforced by olfactory cues from the nearby river, then more hatchlings should have dispersed toward the river. However, 14% moved parallel to or away from water, a proportion that is similar to releases at most other arenas.

Standing et al. (1997) and Tuttle and Carroll (2005) suggested that the variation in orientation of siblings represented bet-hedging or coin-flipping adaptive strategy. However, Pappas et al. (2009) pointed out that it is difficult to develop a scenario where hatchlings dispersing away from wetlands would have higher survival rates than hatchlings moving directly toward wetlands. Examination of dispersal over longer distances and more extended periods would yield information on the ability and success hatchlings may have in making corrections in orientation as they are exposed to new environmental cues.

Other Potential Environmental Cues

The time of day influenced patterns of hatchling snapping turtle orientation at 2 of 3 sites. Regardless of whether there was a time of day change, orientation at all sites was not always toward the sun but appeared to be toward the most illuminated nearby open areas, whether or not those areas were associated with wetlands. The changes in dispersal direction of hatchlings in the morning and afternoon releases at the TNC and DNR Prairie sites were not consistent with orientation based on 1) olfactory cues from nearby wetlands at both sites or 2) specular light from water surface (i.e., at the TNC site the direct view of the wetland was blocked by a band of tall alders).

The results from a study of orientation of yellow mud turtles (Kinosternon flavescens) in 2 arenas on the east and west side of a lake indicate that displaced, experienced hatchlings released in experimental arenas dispersed toward the wetland (Iverson et al. 2009). However, we question the researchers' assertion that an area of higher light intensity caused by reflected light and located over the wetland was the orientation cue. The lake was approximately 200 m away and not visible from the arenas, therefore, hatchlings could not have directly detected reflected nonpolarized or polarized light associated with the wetland. Because hatchling yellow mud turtles were released on different days and at different times, their ability to detect reflected light would have been substantially influenced by the changing position of the sun with respect to the relative locations of the arenas and the wetland. Yellow mud turtle females nest in upland sandhills (Iverson 1990); therefore, positive geotaxis may often be an appropriate water-finding cue. Such a response might have been involved in initial dispersal at the west arena (Iverson et al. 2009), where the direction that hatchlings dispersed (between south and southeast) is downhill (Fig. 2; Iverson et al. 2009). Alternatively, some hatchlings may have dispersed toward the most open, illuminated nearby areas adjacent to the arenas (as do snapping and painted turtles in this study) that were correlated with the direction of the wetland.

Because the terrain inside all of our release arenas was relatively flat or had minimal slope, evidence for or against geotaxis was weak. The results of this study and those of Noble and Breslau (1938) found that hatchling dispersal was nonrandom and apparently based on visual cues. In contrast, Robinson (1989) found that, along the Ottawa River in Quebec, dispersal from nests in flat areas was random and that positive geotaxis increased with slopes from 4°–38°. However, some of the results may have been influenced by 1) steepness of slopes, because hatchling snapping turtles in laboratory experiments lost the ability to climb uphill when slopes were greater than 20° (Noble and Breslau 1938), and 2) correlations between aspect of slopes and the open horizon of the river at nest locations.

Orientation Toward Open Horizons in the Morning and Afternoon

Snapping and painted turtle females nest relatively close to wetlands at Weaver Dunes and at other areas (Hammer 1969; Wilhoft et al. 1979; Obbard and Brooks 1980; Congdon and Gatten 1989; Congdon et al. 2008), and hatchlings of both species oriented and dispersed toward open near horizons (Noble and Breslau 1938; this study). At 1 release arena in 2008 (TNC), hatchling snapping turtles dispersed toward different near open horizons in the morning and afternoon that were not in the direction of a wetland. At the DNR site, direction of dispersal also changed with time of day, with hatchlings moving away from a wetland in morning and toward it in the afternoon. Because the changes in dispersal directions were different at the DNR (east to west) and TNC (south to north-northwest), hatchlings were apparently influenced by changes in the relative illumination of open areas rather than a tendency to move toward the sun. In contrast, at Mallard (an atypical site), where there was no associated wetland, orientation of hatchlings remained toward the same open patch of prairie regardless of time of day. Perhaps the configuration of topographic features resulted in no perceptible change in illumination of the open horizon.

Our results support the insightful study of Noble and Breslau (1938), which indicates that hatchlings primarily used visual cues for orientation toward nearby, open, and most highly illuminated horizons. Our study also indicates that 1) hatchlings disperse toward open horizons rather than toward wetlands themselves (i.e., open areas that are not necessarily associated with wetlands), 2) dispersal direction is influenced by time of day, apparently because of changes in the degree of illumination of different prairie patches, and 3) far horizons were apparently not used because they were beyond the perception distance of hatchlings. Perhaps snapping and painted turtles nest close to wetlands to limit the number of open areas visible to emerging hatchlings and thus increase the probability that the nearest open horizon is associated with a nearby wetland. The variation in patterns of dispersal suggests that, even though snapping and painted turtles nest relatively close to wetlands, problems related to finding wetlands remain complex.

Although we did not make a systematic comparison of hatchling orientation on sunny and overcast days, 3 releases conducted at McCarthy in 2002 (Fig. 6b–d) all resulted in nonrandom dispersal, even though R1 and R2 were on a sunny day and R3 was on a mostly overcast day. Naïve hatchling Blanding's turtles released at the same arena during sunny and overcast days also had nonrandom dispersal on both days (Pappas et al. 2009). The hatchlings were apparently using the local environmental cues and horizon contrasts to orient. In contrast, displaced adult turtles that were apparently using a sun compass had highly directional orientation on sunny days, ceased moving on overcast days (Gould 1957, 1959; Gourley 1974; DeRosa and Taylor 1982; Yeomans 1995; Graham et al. 1996), either because of lower body temperatures or loss of their sun compass.

Comparison of Snapping and Painted Turtles to Blanding's Turtles

The initial orientation of hatchling painted and snapping turtles toward open highly illuminated near horizons is in sharp contrast to hatchling Blanding's turtles that primarily disperse toward dark far horizons (Pappas et al. 2009). Most female painted and snapping turtles leave the wetlands, nest in nearby areas, and return on the same day (Congdon and Gatten 1989; Congdon et al. 2008). Nesting excursions of female Blanding's turtles often are of long duration, and most nests are located much further from the female's home wetland, and further from the nearest wetland than nests of snapping and painted turtles (Congdon et al. 1983, Butler and Graham 1995; Pappas et al. 2000, 2009; but see McNeil et al. 2000).

Traits such as differences in perception distances, morphology, desiccation rates, and risk of being killed or injured by predators combine to influence the nesting tactics of females and the abilities of hatchlings to survive dispersal from nests. For example, snapping turtles desiccate rapidly when on land (Ernst 1968; Finkler 2001), and painted turtles are often killed by predators while in terrestrial nesting areas (J.D. Congdon (2007), unpubl. data). Both species nest close to wetlands and usually take less than half a day to nest. In contrast, the nesting migrations of female Blanding's are much greater in duration and distance than those of snapping and painted turtles, an activity that is presumably facilitated by their ability to close their shells to reduce water loss and lower their risk of death or injury from predators. Because they are small bodied, hatchlings and young juveniles do not have protection that adults obtain from increased body size, and in some cases, plastral kinesis.

Despite differences in locations of arenas in relation to wetlands, the orientations of hatchling snapping and Blanding's turtles were both influenced by time of day. Why either species should alter its response to environmental cues used for orientation at different times of day remains an open question. The relative distances from nests to wetlands are concordant with our estimates of the perception distances of hatchling snapping turtles < 90 m and > 325 m for Blanding's turtles (Pappas et al. 2009). Because orientation and dispersal appears to depend primarily on vision, we predict that 1) the majority of nest-to-wetland distances will be within the species-specific perception distances of hatchlings, and 2) that nesting close to water coupled with hatchling orientation toward nearby open near horizons (painted and snapping turtles) will be the most common case among freshwater turtle species. Perhaps the dichotomy of dispersal cues used by hatchlings dispersing from nests located close to and far from wetlands may have resulted from a decrease in orientation abilities of hatchlings emerging from nests located at intermediate distances from wetlands. If such is the general case, the among-species distribution of nest to wetland distances may turn out to be bimodal rather than continuous.

Although orientation of hatchling freshwater turtles has been studied for only a few species, we think that the most parsimonious explanation for the evolution of orientation and dispersal abilities is that females and hatchlings use the same sensory modalities, perception distances, and environmental cues during terrestrial movements (i.e., females and hatchlings will use the same cues while moving to and from nest sites to wetlands). In addition, the initial orientation abilities of some hatchlings are apparently rapidly enhanced through learning specific landscape cues and development of a sun compass such as found in painted, Blanding's, and mud turtles (DeRosa and Taylor 1982; Caldwell and Nams 2006; Iverson et al. 2009; Pappas et al. 2009).

Hatchling turtles that emerge from nests at a different time of day, on sunny or overcast days, or in different habitats can face substantially different orientation problems. Open prairie nesting areas and large wetlands, such as lakes and rivers, both represent open illuminated horizons; therefore, hatchlings have to discriminate between degrees of openness or subtle differences in horizon contrasts. Wetlands may be the most open illuminated horizon in forested areas, but visual cues associated with wetlands or far dark horizons may often be obscured by nearby upland or riparian trees. Two questions related to variation in emergence conditions need to be addressed: 1) are there hatchling orientation behaviors that overcome orientation problems, or 2) what patterns of variation are associated with orientation errors that increase mortality during dispersal.

Conservation Concerns

Because higher elevation dune soils are less stable, low in organic content, and have low water-holding capacity, the low dune areas that are closest to wetlands are more suitable for agriculture at Weaver Dunes. Both snapping and painted turtles nest close to wetlands and can be directly impacted by agricultural activities in the following ways:

1. Some agricultural activities may prevent females from using previously suitable areas and force them to move farther from water to nest or to use less-suitable nesting areas. If our estimate of perception distance for snapping turtles is reasonably accurate, then females may be forced to construct nests in areas where hatchlings have a higher probability of orienting toward open horizons not associated with wetlands.

2. During the nesting season, when tilled areas are open and crop canopies are small, some females nest in planted agriculture fields that have disturbed soils (i.e., they have very little vegetation cover). However, in fall, dense (soy beans) and tall (corn) crop canopies have developed that may substantially reduce the ability of hatchlings to access suitable environmental cues for orientation.

3. Because painted turtles usually overwinter in nests, machinery used to harvest crops or till soils in fall or early spring may kill hatchlings in nests.

4. Blanding's turtle nests are constructed in higher dunes located relatively long distances from wetlands. Dispersing hatchlings that have to traverse mature corn or soybean fields located between nests and wetlands (or overwintering sites) may have decreased ability to use a sun compass to maintain a direction toward wetlands.

A major conservation issue for the Weaver Dunes turtle community is that a much larger proportion of wetland habitats are protected than are the terrestrial dune areas used for nesting. For each species in the Weaver Dunes turtle community, the major nesting areas should be identified, and potential ways to protect them should be explored. The following conservation measures should increase the probability of successful hatchling recruitment:

1. Use road signs to increase awareness of the presence of hatchlings crossing roads and, perhaps, seasonally reduced speed limits in areas where high densities of hatchlings cross roads (Langen et al. 2008). However, such road signs also identify areas that are good for collecting, therefore, some on-site monitoring may be necessary to counter loss of hatchlings.

2. Remove invasive plants and introduced pine stands that 1) shade nesting areas and degrade the thermal environment for embryo development, and 2) restrict a hatchling's access to appropriate environmental cues for finding overwintering sites and/or wetlands.

3. Where possible, reduce the extent of agricultural fields (and feed plots for game species) in typical turtle nesting areas (or between nesting areas and wetlands) and subsidize farmers to increase the number of fields that lie fallow.