2012 NORTHEASTERN NATURALIST 19(3):361–372 Notes on the Nesting Ecology of Eastern Box Turtles near the Northern Limit of their Range Lisabeth L. Willey1,2,* and Paul R. Sievert2 Abstract - We evaluated the nesting ecology of Terrapene carolina carolina (Eastern Box Turtle) near their northern range limit in Massachusetts. We identifi ed 34 nests in 2005 and 2006 at 4 study sites, and measured clutch size, nest success, hatchling size, and habitat characteristics at each site. Mean clutch size was 5.87 eggs, and egg survival was approximately 50% to hatching, excluding depredation. Large-bodied females tended to oviposit larger clutches than small-bodied females, although the correlation was not signifi cant, and a smaller proportion of their eggs produced live hatchlings. Nest depredation varied greatly across sites from 0 to nearly 100%. The variability observed across the species’ range and across sites underscores the importance of obtaining local information when developing conservation and management programs for rare turtles. The characteristics of the nest sites observed in our study could be simulated to more effectively create or maintain artifi cial nest sites for Eastern Box Turtles in the Northeast. Introduction Although Terrapene carolina carolina L. (Eastern Box Turtle) is well studied throughout much of its range, few studies have examined the species in New England, where it occurs in relatively low densities. Box turtles are protected in most northeastern states in which they occur. In Massachusetts, Eastern Box Turtle is protected as a species of special concern under provisions of the Massachusetts Endangered Species Act (MESA) (M.G.L. Ch. 131A), and the state is currently undergoing a comprehensive conservation planning process for the species (Erb 2011). To inform this and other regional conservation efforts, we evaluated the nesting ecology of Eastern Box Turtles near the northern limit of the species’ range. Eastern Box Turtle nesting ecology has been examined at numerous sites throughout the species’ range (e.g., Allard 1935, Burke and Capitano 2011, Flitz and Mullin 2006, Kipp 2003, Wilson and Ernst 2005), and based on the hypothesis proposed by Iverson (1992) that clutch size increases with latitude, we expected to see larger clutch sizes than those observed in these more southerly studies. Recent work on the relationships among geographic location within a species’ range, abundance, and demographic parameters (e.g., Angert 2009, Gerst et al. 2011) demonstrate the complexity of these relationships, and underscore the importance of assessing variability across species’ ranges and obtaining local demographic parameter values to develop more informed and effective conservation strategies. 1Graduate Program in Organismic and Evolutionary Biology, University of Massachusetts Amherst, North Pleasant Street, Amherst, MA 01003. 2United States Geological Service, Massachusetts Cooperative fish and Wildlife Research Unit, Department of Environmental Conservation, University of Massachusetts Amherst, 160 Holdsworth Way, Amherst, MA 01003. *Corresponding author - lwilley@cns.umass.edu. 362 Northeastern Naturalist Vol. 19, No. 3 Availability of nesting habitat with a suitable thermal regime is one potential factor limiting the distribution of Eastern Box Turtles and other non-marine turtle species (Allard 1935, Bobyn and Brooks 1994, Compton 1999). To address this potential limitation, anthropogenic improvement or creation of nesting habitat for rare turtles is commonly conducted throughout the Northeast (Beaudry et al. 2010, Kiviat et al. 2000, MNHESP 2009), but the nesting habitat requirements for Eastern Box Turtles in the region have not previously been explored. This information is important for developing such site-specifi c management plans and to assist the state-wide conservation planning process. We assessed nesting habitat, clutch size, and nest success by radio-tracking female box turtles at 4 study sites in 2005 and 2006. We also related clutch size and nest success to female body size and habitat characteristics. field-Site Description This study was conducted at 4 study sites in the Connecticut River Valley, MA. Study sites were selected across a range of human land-use intensity, habitat type, latitude, and elevation (Table 1). Sites were distributed throughout the Valley along a north–south gradient of 56 km and ranged in elevation from 50 m to 300 m. The use of multiple sites with different habitat characteristics provides an opportunity to estimate variation in nesting parameters. To quantify habitat differences between sites, the percentage of forested, open canopy, and developed area within each site was measured using 1999 land-use cover data from the Offi ce of Geographic and Environmental Information (MassGIS), Commonwealth of Massachusetts Executive Offi ce of Energy and Environmental Affairs. The 4 sites are described below. The exact names and localities of the sites are withheld for conservation purposes. Site A (50 ha) is a small, municipally owned conservation area that consists of Acer rubrum L. (Red Maple)-dominated forest, a wetland complex, and a former gravel-extraction site. This site is surrounded by active agriculture and residential development and is heavily used by hikers. Site B (100 ha) is a parcel of private conservation land consisting of a pineoak upland and a beaver-dammed brook flowing through the low-lying portion of the site. The adjacent lot was developed in 2004–2005, and onsite mitigation Table 1. Summary of Eastern Box Turtle study sites. Site Area % % open % name (ha)A forestedB Dominant tree species habitatB developedB Site A 50 25% Acer rubrum 32% 43% Site B 100 78% Pinus strobus, Quercus spp. 11% 11% Site C 900 79% Pinus rigida, Quercus spp., 16% 5% Acer saccharum Marshall (Sugar Maple) Site D 1200 90% Quercus species, Carya spp., Tsuga 5% 4% canadensis (L.) Carrière (Eastern Hemlock) ACalculated as the continuous land area to nearest paved road, used as an approximation for human land-use intensity. BCalculated as percent of area in specifi ed land use using MassGIS 1999 land-use data. 2012 L.L. Willey and P.R. Sievert 363 for box turtles was required under MESA. The development was surrounded by a “turtle curb”, which consisted of a continuous concrete barrier greater than 30 cm in height, designed to confi ne the animals to the undeveloped area. An artifi cial nesting area was also constructed within the Eastern Box Turtle habitat. Within the last 20 years, the site has been surrounded by development including residential neighborhoods, schools, and light industrial uses, primarily shipping and receiving. Site C (900 ha) is a state-owned conservation area where prescribed burns are used for restoration and habitat maintenance. The site is used by the public for recreation (e.g., hunting, walking, jogging, and all-terrain vehicles), and consists of a deciduous forested hillside adjacent to a large tract of Pinus rigida Mill. (Pitch Pine)-scrub oak barrens with areas of open sand. Site D (1200 ha) comprises the south slope of an east–west-trending basaltic mountain range (up to about 300 m in elevation). This site has little human use (hiking, hunting, and some all-terrain vehicle use), and it is bordered by residential development, a major state highway, and gravel operations to the east, south, and west. This site is primarily composed of private forest land, but is bisected by an east–west treeless right-of-way. Methods Turtles were captured using visual encounter surveys. We individually marked each turtle by fi ling the marginal scutes (Ernst et al. 1974). We then measured each animal using dial calipers, and photographed it in the fi eld. For each turtle, we measured straight carapace length (SCL), total carapace length (TCL), plastron width at the humeral/pectoral seam (PW), carapace width at the widest point (CW), and carapace height at the deepest point (CH). We categorized turtles into 6 age classes by counting the lines of arrested growth (LAG) on their plastron and evaluating relative amounts of shell-wear on the plastron: Class 1 = visible new growth on plastron, no wear; Class 2 = no new visible growth, no wear; Class 3 = beginning to wear; Class 4 = less than 50% worn; Class 5 = more than 50% worn; Class 6 = plastron worn smooth. At each site, a subset of turtles (5–11 adult females) was outfi tted with radio transmitters (MBFT-6, Lotek Wireless, Newmarket, ON, Canada; R2020, Advanced Telemetry Systems, Isanti, MN). We affi xed radios along the posterior margin of the carapace using dental acrylic (Biocryl Resin, Great Lakes Orthodotics, Tonawanda, NY). For MBFT-6 models, antennae were affi xed with dental acrylic to the carapace in a ring along the pleural/marginal seam (as in Compton 1999); for R2020 models, the antennae were shorter and were free to trail behind the turtle. Total weight of radio and acrylic was less than 5% of body weight. Turtles were released immediately after processing. We located turtles using radio-telemetry 2–3 times per week throughout the spring and May-to-June nesting season and recorded positions to the nearest 5 m using a hand-held global positioning system (GPS) receiver (eTrex 12-channel or GPSmap 76CSx, Garmin International, Inc., Olathe, KS). As radio-equipped females neared potential nest sites, we located them more frequently (up to twice per 364 Northeastern Naturalist Vol. 19, No. 3 day) and weighed and palpated them to determine if they were gravid. If females were gravid, we affi xed a thread bobbin to the caudal portion of the carapace (using the method described in Milam 1997). When we located the animal in the morning, we assessed whether she remained gravid. If she was no longer gravid, or if we were unable to palpate her but she had lost considerable weight from the previous night, we followed her thread trail to locate the nest chamber (Milam 1997). Potential nest sites were indicated by an erratic thread trail, disturbed soil, or areas where the thread was buried. We collected the thread and removed it from the site when we located the chamber, or if it could not be located after substantial searching. We covered nests with half-inch-mesh hardware-cloth screens to protect eggs from predators. We replaced screens with hardware-cloth box screens in August to allow hatchlings to emerge without escaping (Graham 1997), and later checked nests daily for emerged hatchlings. We measured and released the hatchlings upon emergence. Nests were excavated in November, after hatchlings had emerged, in order to determine clutch size and nest success. Success was measured as the proportion of eggs that developed and successfully emerged from the nest. At each nest site, we collected environmental data, including cover type, substrate type, distance to nearest ecotone, and vegetation structure within a 5-m radius of the nest. Within that circle, we visually approximated percent canopy (greater than 3 m in height), percent shrub cover (woody stems less than 3 m in height), percent herbaceous cover, leaf litter, and bare ground. We also recorded the 3 dominant plant species of each layer. We measured the dimensions of the canopy openings used by turtles for nesting using 2005 orthophotos (MassGIS) and ArcMap 9.2 (ESRI, Redlands CA). We also measured the area of the canopy opening and the distance of each nest to the nearest forest edge in the four cardinal directions via GIS. We evaluated the effect of site and calendar year on clutch size and nest success using ANOVA. We used linear regression to explore the effects of age, body size, and habitat characteristics on clutch size and nest success. To maintain independence of samples, only one nest per female was included in the analysis. The fi rst (2005) nest of females with nests in both years was used. This procedure resulted in a sample size of 24 for ANOVA and regression analyses. Count values (i.e., age class, clutch size, and number hatched) were square root transformed, and proportions (i.e., forest cover and nest success rate) were arcsine transformed. Residuals were visually assessed to ensure assumptions were not violated. Analyses were conducted in R (R development core team 2010). Results We screened 34 Eastern Box Turtle nests at 4 sites in 2005 and 2006. Nesting was observed in the Connecticut River Valley from 27 May to 10 July, with the peak occurring in early June. The fewest LAGs exhibited on any female that we observed nesting was 14 LAGs, and her SCL was 144 mm. We tracked 3 females with 12 and 13 LAGs (119–140 mm SCL), but these animals did not nest. The smallest-bodied female that we observed nesting was 122 mm SCL, but her carapace was entirely worn smooth, and she was probably very old. 2012 L.L. Willey and P.R. Sievert 365 Approximately 90% of mature females nested annually. Turtles were carefully monitored for evidence of double clutching, and no instances were documented. Of the 34 nests, two were depredated despite protective screens, and one was destroyed by mechanized equipment. Results from the remaining 31 nests are presented in Table 2. Clutch size ranged from 3 to 10 (mean = 5.87, sd = 1.88). Twenty-four of the 31 nests produced at least one hatchling, and mean nest success (average proportion of eggs hatching in each nest), excluding depredation or human disturbance, was 53%. Of the 182 eggs oviposited in all 31 nests, 55% hatched successfully, 28% failed to develop, 16% died prior to hatching, and 1% hatched but died in the nest chamber. In 2005 and 2006, hatchlings emerged between 20 August and 9 October. Upon emergence, hatchling carapace length ranged from 26.7 to 35.2 mm (mean = 31.9, n = 48). Average hatchling size was not correlated with the body size of the mother (F1,12 = 0.00, P = 0.94). The hatchlings from larger clutches tended to be smaller in size, but not signifi cantly (F1,12 = 2.4, P = 0.14). Correlation between body size, age, habitat and reproductive rates Between sites, there were no signifi cant differences in clutch size (F3,20 = 1.3, P = 0.31), hatching number (F3,20 = 0.94, P = 0.44), or success rate (F3,20 = 1.4, P = 0.26). Similarly, there were no signifi cant differences between years (F3,20 < 0.14, P > 0.7 in all cases), and therefore sites and years were pooled for the remaining analyses. Female body size was positively correlated with the square root of clutch size (fig. 1A), and TCL was the body size metric best explaining clutch size, although this relationship was not signifi cant (F1,22 = 1.98, P = 0.17). Carapace height showed no linear relationship with transformed clutch size (F1,22 = 0.00, P = 0.95); the three females that deposited the largest clutches (9 or 10 eggs) had carapace heights (68–71 mm) close to the mean (68 mm) (fig. 1B). Transformed nest hatching rate was negatively correlated with total carapace length (F1,22 = 5.0, P = 0.04) and carapace height (F1,22 = 4.5, P = 0.04). Larger females had more frequent nest failures than smaller-bodied females (fig. 1D, E), but because larger females oviposit larger clutches, the total number of successful hatchlings was not signifi cantly correlated with the body size of the mother (F1,22 = 2.6, P = 0.12). The mother’s age class, ranging from 1 to 6, had no signifi cant linear relationship with transformed clutch size (F1,22 = 1.07, P = 0.31) or nest success (F1,22 = 0.88, P = 0.36) (fig. 1C, F). Clutch size and nest success were not signifi cantly correlated with any habitat variable (F1,22 < 1.47, P > 0.23 in all cases). Turtles in heavily forested areas Table 2. Summary statistics for the size and success rate of 31 box turtle nests observed in the Connecticut River Valley, MA. Analysis excludes nests that were depredated or destroyed. Clutch size Successfully emerged hatchlings Success rate Minimum 3 0 0.00 Median 6 3 0.67 Mean 5.87 3.23 0.53 Maximum 10 8 1.00 Standard deviation 1.88 2.49 0.35 366 Northeastern Naturalist Vol. 19, No. 3 tended to lay smaller clutches, whereas those in more open habitats deposited larger, less successful clutches, although these relationships were not signifi cant (F1,22 = 1.47, P = 0.24). Depredation Although we screened nests of radio-tagged turtles and therefore did not directly measure the effects of depredation on box turtle nests, 2–6 incidentally observed box turtle nests at each site were not screened. The effect of depredation on nests that were not screened appeared to vary considerably. Although sample size was low, sites with higher nest density and a greater number of turtle species present that were located in areas surrounded by anthropogenic land cover had higher depredation rates (up to 100%). All 4 nests were completely destroyed by predators at the site (Site A) that is surrounded by residential and agricultural land uses. Two of these nests were screened and 2 were not. Six unscreened nests located in a remote power line right-of-way through a forested area (Site D) were left completely intact (0% depredation). Nest-site characterization Turtles used abandoned gravel pits, right-of-ways, backyards, old fields, and forest clearings as nest sites. Nest sites tended to be sandy, open areas with little vegetation, but all nests were deposited within 1 m of vegetation. figure 1. Eastern Box Turtle clutch size and nest success plotted against body size. Larger females tend to lay larger clutches (though not signifi cantly), and they are signifi cantly less successful than those oviposited by smaller bodied females. 2012 L.L. Willey and P.R. Sievert 367 Canopy cover within 5 m of nests (n = 34) ranged from 0 to 65% (median = 0), shrub cover within 5 m ranged from 0 to 40% (median = 5%), and herbaceous cover within 5 m ranged from 5% to 90% (median = 45%). Quercus spp. (oak) dominated the canopy and shrub layers whereas various species of graminoids, Solidago spp. (goldenrod), and Potentilla spp. (cinquefoil) were the most commonly occurring herbaceous plants around the nest. At nest sites, turtles spent more time near the nest area if woody or herbaceous material in which to hide and forage was present. Areas with downed wood, graminoids, goldenrod species, Comptonia peregrina (L.) Coult. (Sweet Fern), Rubus spp. (blackberry/raspberry/dewberry), Rhus typhina L. (Staghorn Sumac), Rosa multiflora Thunberg (Multiflora Rose), Lonicera spp. (Honeysuckle), and Elaeagnus umbellata Thunberg (Autumn Olive), as well as young Betula spp. (birch), Populus spp. (aspen), and oak species were frequently used for resting and foraging by females immediately before and after nesting. Nest-site open-canopy dimensions There were 16 nesting areas at the 4 study sites. Median dimensions of clearings used by nesting turtles were 125 m on the north–south axis and 54 m on the east–west axis. Nest site openings were generally longer on the north–south axis. Minimum axes were 19 and 7 m, respectively. The minimum observed opening used as a nest site was approximately 1200 m2, and the median was 5670 m2 (Table 3). Table 4 lists the distances of observed turtle nests from various forest edges. All of the nests we observed were closer to the northern edge of sandy openings than to the southern edge. Table 3. Dimensions of canopy openings used by Eastern Box Turtles for nesting. Dimension Minimum Median Maximum North–south axis length (m) 19 125 1000A East–west axis length (m) 7 54 377 Approximate clearing area (m2) 1200 5670 72,384 AFor clearings that are very long (e.g., power-line corridors), a maximum dimension of 1000 m was used. Table 4. Distances (m) from Eastern Box Turtle nests to forest edges. Minimum 5th percentile Median Maximum North 0 0.0 12 500A Northeast 0 3.2 16 100 East 0 2.4 27 78 Southeast 14 14.0 35 117 South 10 22.0 125 500A Southwest 12 15.2 43 175 West 0 3.2 27 361 Northwest 0 0.8 20 52 AFor clearings that are very long (e.g., power-line corridors), a maximum distance of 500 m was used. 368 Northeastern Naturalist Vol. 19, No. 3 Discussion Clutch size and nest success The clutch size we observed was larger than that reported from most other studies (Table 5). There was no evidence of double clutching in our sample, although this has been reported fairly regularly farther south; double clutching was also not observed in Long Island (Cook 2004). Our results support the observation, generally suggested in the literature (e.g., Iverson 1992, Iverson et al. 1993, Wilson and Ernst 2005) that clutch size increases with latitude, whereas clutch frequency decreases with latitude. The proportion of females nesting in a year (90% in our sample) was higher than that observed in other studies (Cook 2004, Dodd 2001, Wilson and Ernst 2005). Nest success rate (55%), which excluded depredation, was similar to that reported in other studies (e.g., Belzer 2002, Kipp 2003, Wilson and Ernst 2005). However, the methodology among studies in estimating success varied, so it is diffi cult to directly compare predation rates. While it is possible that human intervention (i.e., fi nding and screening the nest) lowered success rates, our methods were similar and in many cases less invasive than those used in other studies. In addition, Samson et al. (2007) found that handling did not decrease the rate of nest productivity in Chrysemys spp. (painted turtles), suggesting a natural mechanism for the low success observed in our case. Because nesting habitat with the appropriate thermal regime for successful incubation is thought to be one factor limiting turtles at the northern limit of their range (Allard 1935, Bobyn and Brooks 1994, Compton 1999), low success rates are not surprising and suggest that any advantage conferred by larger clutch sizes in northern areas may offset lower success rates. Limited nesting habitat availability and incubation temperature suitability in Massachusetts may explain why box turtles tend to be distributed in the warmer, sandier portions of the state: Cape Cod, the southeast, and the Connecticut River Valley. Future work on the thermal regime of nests near the northern edge of the species range may help elucidate this question. Variation in both clutch size and success rate across the species’ range underscores the importance of adjusting parameters to local demographics when Table 5. Clutch sizes of Eastern Box Turtles reported throughout their range. Author Location Average clutch size Allard 1935 Washington, DC 4.2 Ewing 1935 Washington, DC 3.8 Congdon and Gibbons 1985 South Carolina 3.4 Stuart and Miller 1987 North Carolina 3.0 Mitchell 1994 Virginia 4.1 Tucker 1999 Illinois 4.9 Belzer 2002 Pennsylvania 4.0 Kipp 2003 Delaware 4.6 Cook 2004 Long Island, NY 5.8 Wilson and Ernst 2005 Virginia 3.15 Burke and Capitano 2011 Long Island, NY 4.1 This study Massachusetts 5.87 2012 L.L. Willey and P.R. Sievert 369 assessing status and developing conservation plans. For instance, a populationviability analysis utilizing the larger northern clutch size with a higher observed nest success rate from another regional study would be positively biased with regard to recruitment and population growth. Conversely, using smaller clutch sizes reported for more southern populations with mortality rates from a site with lower nest success rates would lead to exaggerated projections of population decline. Body size Although the positive correlation between clutch size and body size has been reported in a number of studies (Iverson 1992, Kipp 2003, Tucker 1999, Wilson and Ernst 2005), few studies have evaluated the correlation between the mother’s body size and number of successful hatchlings. Our results suggest that in clutches deposited by larger mothers, a smaller proportion of eggs successfully hatch, potentially offsetting the benefi t conferred by larger clutch size, although larger clutches still produced signifi cantly more hatchlings (F1,22 = 4.62, P = 0.04). Although smaller females tended to produce a greater proportion of successful eggs, because there was no correlation between female body size and the total number of hatchlings or hatchling size, and because we did not follow hatchlings beyond emergence, we do not know the associated recruitment rates or whether larger- or smaller-bodied turtles are more likely to produce offspring that survive to maturity. Allard (1935) and Tucker (1999) found similar results whereby large female Eastern Box Turtles deposited larger clutches, but large clutches contained eggs that were smaller and weighed less, generally agreeing with the pattern discussed by Iverson et al. (1993). We did not measure or weigh eggs in our sample, so we do not know whether there was a correlation between egg size and hatching rate. Depredation High depredation rates of turtle nests are frequently reported. For example, in Illinois, Flitz and Mullin (2006) observed 87.5% depredation on Eastern Box Turtle nests that were not protected. However, predation rates are difficult to measure because they vary substantially across both space and time with weather, predator densities, landscape characteristics, etc. (e.g., Bowen and Janzen 2005, Kolbe and Janzen 2002, Strickland and Janzen 2010), and it is possible that observer disturbance increases susceptibility to nest predation (Rollinson and Brooks 2007). We found that depredation rates varied greatly by location, similar to variation previously reported in Eastern Box Turtles (e.g., Kipp 2003). While our sample size was not large enough to draw statistical conclusions about the relationship between habitat variables and nest depredation rates, depredation rates on nests are likely a function of local predator density as well as turtle nest density. In other systems, edges have been shown to increase the prevalence of meso-predators (Dijak and Thompson 2000, Herkert et al. 2003), and in some cases, turtle nests near edges have been shown to be more susceptible to depredation (Strickland and Janzen 2010, Temple 1987), but this effect has not been consistent across studies (Kolbe and Janzen 2002, Marchand and Litvaitis 2004). 370 Northeastern Naturalist Vol. 19, No. 3 The 3 nests of the original 34 nests that were destroyed by predators and humans were located at sites with the greatest amount of human activity. Both sites were surrounded by residential development. Site A, the site with depredation rates up to 100%, is a residential and agricultural area that may support larger numbers of meso-carnivores. Chrysemys picta picta Schneider (Eastern Painted Turtle) and Chelydra serpentina serpentina L. (Eastern Snapping Turtle) use the same nesting area as Eastern Box Turtles at this site, creating turtle nest densities that were higher than those at other sites (L.L. Willey, pers. observ.). Site D, which had very low depredation, is mostly forested, and no other species of turtle has been observed there. While limited, our observations seem to support the conclusion of previous authors (e.g., Marchand and Litvaitis 2004) who observed that depredation was higher in areas with high nest density and near agricultural sites. Habitat management implications Clearing land to open up habitats for nesting turtles is conducted throughout Massachusetts by both MassWildlife and USDA Natural Resources Conservation Service (MNHESP 2009; B. Schreier, USDA Natural Resources Conservation Service, Amherst, MA, pers. comm.) and throughout the Northeast (Kiviat et al. 2000). Our results can be used to inform the design of such management and suggest that canopy openings should be at least 1200 m2 and probably larger to attract nesting box turtles. Openings where turtles nested were generally longer on the north–south axis than the east–west axis. This fi nding is probably related to insolation. Interestingly, females that inhabited east–west oriented powerline habitat prior to nesting in May sometimes traveled up to 1200 m to nest elsewhere, whereas those that used north–south oriented power-lines generally nested at that location. Sample size was small, so other attributes (e.g., soil or vegetation type) might have influenced nest-site choice. We would still recommend that nesting habitat should be created on a north–south axis. Beaudry et al. (2010) reported that adult Emydoidea blandingii Holbrook (Blanding’s Turtle) are willing to use anthropogenically altered areas, but there is no empirical evidence to evaluate nesting success in such sites. Future studies should address the frequency and viability of nests in managed areas. Ideally, management plans should be developed with all life stages in mind, but particular precaution should be taken to avoid adult mortality during management activities. Acknowledgments We thank the Massachusetts Natural Heritage and Endangered Species Program, The University of Massachusetts Graduate Program in Organismic and Evolutionary Biology, the Department of Natural Resources Conservation, the USGS Massachusetts Cooperative fish and Wildlife Research Unit, The University of Massachusetts Natural History Collections, and the Turtle Conservation Project for funding support. We also thank M. Jones, L. Johnson, Z. Dowling, D. Yorks, C. Jordan, B. Dunphy, and B. Crowley for their help in the fi eld, and A. Breisch, R. Cook, S. Fowle, C. Griffi n, M. Jones, K. McGargial, A. Richmond, and two anonymous reviewers for providing helpful comments on previous versions of this manuscript. We thank the many landowners for use of their property during the course of this study. Methods were approved by the University of Massachusetts at Amherst Institutional Animal Care and Use Committee (protocol # 25-02-04). 2012 L.L. Willey and P.R. Sievert 371 Literature Cited Allard, H.A.1935. The natural history of the box turtle. Scientifi c Monthly 41:325–338. Angert, A.L. 2009. Biogeography, changing climates, and niche evolution Sackler colloquium: The niche, limits to species’ distributions, and spatiotemporal variation in demography across the elevation ranges of two monkeyflowers. Proceedings of the National Academy of Science of the USA 106:11641–11645. Beaudry, F., P.G. deMaynadier, and M.L. Hunter. 2010. Nesting movements and the use of anthropogenic sites by Spotted Turtles (Clemmys guttata) and Blanding’s Turtles (Emydoidea Blandingii). Herpetological Conservation and Biology 5:1–8. Belzer, B. 2002. A nine-year study of Eastern Box Turtle courtship with implications for reproductive success and conservation in a translocated population. Turtle and Tortoise Newsletter 6:17–26. Bobyn, M.L., and R. Brooks. 1994. Incubation conditions as potential factors limiting the northern distribution of Snapping Turtles, Chelydra serpentina. Canadian Journal of Zoology 71:28–37. Bowen, K.D., and F.J. Janzen 2005. Rainfall and depredation of nests of the Painted Turtle, Chrysemys picta. Journal of Herpetology 39(4):649–652. Burke, R.L., and W. Capitano. 2011. Nesting ecology and hatching success of the Eastern Box Turtle, Terrapene carolina, on Long Island, New York. American Midland Naturalist 165(1):137–142. Compton, B. 1999. Ecology and conservation of Wood Turtles in western Maine. M.Sc. Thesis. University of Maine, Orono, ME. Congdon, J.D., and J.W. Gibbons 1985. Egg components and reproductive characteristics of turtles: Relationships to body size. Herpetologica 41(2):194–205. Cook, R.P. 2004. Dispersal, home-range establishment, survival, and reproduction of translocated Eastern Box Turtles, Terrapene c. carolina. Applied Herpetology 1:197–228. Dijak, W.D., and F.R. Thompson III. 2000. Landscape and edge effects of the distribution of mammalian predators in Missouri. Journal of Wildlife Management 64:209–216. Dodd, K. 2001. North American Box Turtles: A Natural History. University of Oklahoma Press, Norman, OK. 231 pp. Erb, L. 2011. Eastern Box turtle conservation plan for Massachusetts. Massachusetts Division of fisheries and Wildlife, Massachusetts Natural Heritage and Endangered Species Program. Available online at http://www.mass.gov/dfwele/dfw/nhesp/conservation/ herps/pdf/teca_cons_plan.pdf. Accessed 19 June 2011. Ernst, C.H., M.F. Hershey, and R.W. Barbour, 1974. A new coding system for hardshelled turtles. Transactions of the Kentucky Academy of Science 35:27–28. Ewing, H.E. 1935. Further notes on the reproduction of the Eastern Box Turtle, Terrapene carolina (Linne). Copeia 1935:102–103. Flitz, B.A., and S.J. Mullin. 2006. Nest-site selection in the Eastern Box Turtle (Terrapene carolina carolina) in Illinois. Chelonian Conservation and Biology 5(2):309–312. Gerst, K.L., A.L. Angert, and D.L. Venable. 2011. The effect of geographic range position on demographic variability in annual plants. Journal of Ecology 99:591–599. Graham, T.E. 1997. Effective predator excluders for turtle nests. Herpetological Review 28:76. Herkert, J.R., D.L. Reinking, D.A. Wiedenfeld, M. Winter, J.L. Zimmerman, W.E. Jensen, E.J. fincke, R.R. Koford, D.H. Wolfe, S.K. Sherrod, M.A. Jenkins, J. Faaborg, and S.K. Robinson. 2003. Effects of prairie Fragmentation on the nesting success of breeding birds in the midcontinental United States. Conservation Biology 17:587–594. 372 Northeastern Naturalist Vol. 19, No. 3 Iverson, J.B. 1992. Correlates of reproductive output in turtles (Order Testudines). Herpetological Monographs 6:25–42. Iverson, J.B., C.P. Balgooyen, K.K. Byrd, and K.K. Lyddan. 1993, Latitudinal variation in egg and clutch size in turtles. Canadian Journal of Zoology 71:2448–2461. Kipp, R.L. 2003. Nesting ecology of the Eastern Box Turtle (Terrapene carolina carolina) in a fragmented landscape. M.Sc. Thesis. University of Delaware, Newark, DE. Kiviat, E., G. Stevens, R. Brauman, S. Hoeger, P.J. Petokas, and G.G. Hollands. 2000. Restoration of wetland and upland habitat for the Blanding’s Turtle, Emydoidea blandingii. Chelonian Conservation and Biology 3:650–657. Kolbe, J.J., and F.J. Janzen. 2002. Spatial and temporal dynamics of turtle nest predation: Edge effects. Oikos 99:538–544. Marchand, M.N., and J.A. Litvaitis. 2004. Effects of landscape composition, habitat features, and nest distribution on predation rates of simulated turtle nests. Biological Conservation 117:243–251. Massachusetts Natural Heritage and Endangered Species Program (MNHESP). 2009. Advisory guidelines for creating turtle nesting habitat. Massachusetts Division of fisheries and Wildlife, Westborough, MA. Available online at http://www.mass.gov/ dfwele/dfw/nhesp/conservation/pdf/creating_turtle_nesting_sites.pdf. Accessed 29 July 2010. Milam, J. 1997. Ecology and conservation of Spotted Turtles, Clemmys guttata, in central Massachusetts. M.Sc. Thesis. University of Massachusetts, Amherst, MA. Mitchell, J.C. 1994. The Reptiles of Virginia. Smithsonian Institution Press, Washington, DC. 352 pp. R Development Core Team. 2010. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at http://www.R–project.org. Rollinson, N., and R.J. Brooks. 2007. Marking nests increases the frequency of nest depredation in a northern population of Painted Turtles (Chrysemys picta). Journal of Herpetology 41(1):174–176. Samson, J., E.J. Hughes, and R.J. Brooks. 2007. Excavation is a nondeleterious method for obtaining fecundity and morphometric data from small-sized eggs of freshwater turtles. Chelonian Conservation and Biology 6(2):255–259. Strickland, J.T., and F.J. Janzen. 2010. Impacts of anthropogenic structures on predation of Painted Turtle (Chrysemys picta) nests. Chelonian Conservation Biology 9(1):131–135. Stuart, M.D., and G.C. Miller. 1987. The Eastern Box Turtle, Terrapene c. carolina (Testudines: Emydidae), in North Carolina. Brimleyana 13:123–131. Temple, S.A. 1987. Predation on turtle nests increases near ecological edges. Copeia 87:250–252. Tucker, J. K. 1999. Reproductive output of Terrapene carolina, Chrysemys picta, and Sternotherus odoratus from west-central Illinois. Bulletin of Maryland Herpetological Society 35(3):61–75. Wilson, G.L., and C.H. Ernst. 2005. Reproductive Ecology of Terrapene carolina carolina (Eastern Box Turtle) in Central Virginia. Southeastern Naturalist 4(4):689–702.