2007 SOUTHEASTERN NATURALIST 6(4):715–726 Use of Altered Habitats by the Endemic Sand Skink (Plestiodon reynoldsi Stejneger) David A. Pike1,2,*, Kelley S. Peterman1, and Jay H. Exum1 Abstract - Endemic species with limited geographic ranges are particularly sensitive to habitat loss and degradation. In order to conserve such species, we tend to focus on optimal habitats (e.g., for land acquisition), but human-altered habitats may also be of conservation value if species can persist in these environments. We evaluated the occurrence of a lizard endemic to scrub and sandhill habitats in Florida, the threatened Plestiodon reynoldsi (Sand Skink), to determine whether it occurs in human-altered vegetation types. From 2003–2004, we quantitatively sampled 46.5 ha (composed of 7 vegetation and 10 soil types) for the distinctive trails this fossorial lizard leaves in sandy areas. The Sand Skink was present in equal relative densities across all vegetation types, including disturbed areas such as pastures and abandoned citrus groves. Further, we detected Sand Skinks on all well-drained sandy soils we sampled. Thus, Sand Skinks can persist in humanaltered habitats, at least when underlying soils are suitable for their presence and have not been modified. In general, more Sand Skinks were found in larger habitat patches, regardless of whether patches were characterized by vegetation or soil. Our results suggest that anthropogenically altered habitats have conservation value for Sand Skinks (and possibly other ecologically similar species) and that future studies should focus on the effects of restoring these habitats on resident populations of this species. Introduction Many species are declining due to habitat loss and fragmentation (Schneider 2001), and current research and protection efforts often focus on remaining high-quality habitats (e.g., McCoy et al. 1999). However, studying species in optimal habitats alone may not provide sufficient information to determine which specific environmental characteristics affect species presence. Historically, natural disturbance events maintained vegetative structure in many habitats. However, increasing habitat fragmentation has suppressed disturbances, resulting in overgrown habitats that may be unsuitable for many species (e.g., Pringle et al. 2003, Webb et al. 2005). Understanding which habitat changes are occurring and how species react to these changes provides important information about ways to protect biodiversity (e.g., Jäggi and Bauer 1999, Webb and Shine 1998, Webb et al. 2005). For example, past research proposed that habitat- specific endemic lizards (e.g., Plestiodon reynoldsi Stejneger [Sand 1Environmental Services Group, Glatting Jackson Kercher Anglin Inc., 120 North Orange Avenue, Orlando, fl32801. 2Current address - School of Biological Sciences A08, University of Sydney, NSW 2006, Australia. *Corresponding author - dpike@ student.usyd.edu.au. 716 Southeastern Naturalist Vol. 6, No. 4 Skink], Scincidae) might not require habitats containing high-quality natural vegetation to persist (Campbell and Christman 1982). Instead, the abundance and distribution of this fossorial, sand-swimming lizard may largely be determined by soil characteristics because locomotion and movement of this species is impeded by compacted soils (Andrews 1994, Lee 1969). If human-altered habitats harbor important populations of habitat-specific endemic species (e.g., Panzer et al. 1995), protection and restoration of these areas may aid in species persistence (Ries et al. 2001). Conversely, if such habitats are ignored, naturally occurring and important populations may be destroyed. Without a thorough understanding of all factors restricting a species’ distribution, important populations may go unnoticed. For example, habitats are often characterized by vegetation type, even though other factors such as elevation, rainfall, soil type, or thermal environments may be largely responsible for species presence (Campbell and Christman 1982, Webb and Shine 1998, Webb et al. 2005). In fact, a combination of habitat characteristics (aside from simple vegetation types) may be what shapes species distributions, especially on small spatial scales and in severely fragmented habitats (Webb and Shine 1998). We studied the habitat associations of Plestiodon (formerly Neoseps [Schmitz et al. 2004]) reynoldsi, a lizard restricted to central Florida that is locally abundant in scrub and sandhill habitats, which are habitats frequently subject to some degree of human disturbance. Despite this, little is understood about local-scale distribution patterns. To help understand the local-scale distribution of the Sand Skink, we surveyed altered habitat types and focused on the influence of vegetation type on relative density of Sand Skinks. As ectotherms, lizards rely on many aspects of the habitat for thermoregulation, which can affect foraging efficiency, digestion, growth, and egg development. For example, structural aspects of the vegetation heavily influence light penetration, resulting in direct effects on thermal profiles (Pringle et al. 2003, Webb et al. 2005). Altered habitats may contain the habitat characteristics (e.g., structure) that provide adequate thermal regimes for Sand Skinks. Additionally, because Sand Skinks are fossorial, they may require soil types that allow subsurface movements, particularly loose, dry soils (Campbell and Christman 1982). We also studied the relationship between soil type and Sand Skink presence, irrespective of the vegetation type present. Materials and Methods Study species Sand Skinks are small fossorial lizards (≈57 mm snout–vent length; Ashton 2005) that “swim” just beneath the surface of the soil. This swimming motion leaves distinct sinusoidal trails on the surface of the sand. 2007 D.A. Pike, K.S. Peterman, and J.H. Exum 717 Because Sand Skinks spend much of their time belowground, these trails are used as a surrogate for animal presence (Andrews 1994, Sutton et al. 1999). Several other declining or habitat-restricted reptile species also occur sympatrically (Branch et al. 2003, Campbell and Christman 1982, Means and Simberloff 1987), including P. [Eumeces] egregius lividus Mount (Blue-tailed Mole Skink), another elusive species of semi-fossorial skink that is of conservation concern (United States Fish and Wildlife Service 1993). The conservation of Sand Skinks and other sympatric, imperiled species is critical to maintaining an important segment of the highly diverse reptile fauna of the southeastern United States (Gibbons and Stangel 1999). Habitat associations of Sand Skinks Several broad-scale habitat characteristics have been emphasized as supporting Sand Skink populations. The overall geographic distribution is largely defined by relatively high elevations (>30 m above sea level) occurring along relict sand dunes remaining from periods of elevated sea levels during the Pleistocene (Telford 1959). Associated habitats include rosemary scrub, overgrown scrub, high pine, and longleaf pine-turkey oak sandhills (Campbell and Christman 1982; Cooper 1953; Telford 1959, 1962). Ecotonal areas are also important, particularly those between rosemary scrub and palmetto-pine flatwoods habitats, and between longleaf pine and sand pine scrub habitats (Cooper 1953; McCoy et al. 1999; Telford 1959, 1962). Within these habitats, Sand Skinks are often associated with open, sandy patches that allow subsurface movement (e.g., free of roots and grasses that may impede fossorial locomotion), but are most often found underneath cover (Cooper 1953, Telford 1959). Campbell and Christman (1982) also hypothesized that reptiles in scrub and sandhill habitats are distributed in relation to physical characteristics of the habitat itself, rather than to specific vegetation types. Therefore, areas that contain well-drained, sandy soils and open patches free of root mats may be suitable for skinks (Campbell and Christman 1982), even though surface vegetation types are disturbed or converted to agricultural production, particularly citrus. Study site We studied skinks on a 158-ha tract of private land in Lake County, FL. The site is composed of several upland habitat types growing on well-drained, sandy soils (Table 1, Fig. 1). Historic cattle ranching and the production of citrus fruit have altered the natural vegetation, which is now mostly characterized as early stage successional systems (Table 1). Little natural vegetation of the kind normally associated with natural Sand Skink habitat remains due to these historical habitat alterations, and the habitat surrounding our site is extensively modified, and much of it is currently being developed. Thus, little (if any) habitat typically thought of as appropriate for Sand Skink exists in the area. 718 Southeastern Naturalist Vol. 6, No. 4 Sampling regime We established 9 sampling plots of varying sizes (mean = 5.18 ± 1.84 ha; range = 0.43–13.58 ha) in vegetation and soil types representative of our site (totaling 46.5 ha; Fig. 1). Plots were based on representative vegetation types and avoided wetland areas, resulting in irregular sizes and shapes Table 1. A description of the vegetation types at our study site in Florida that are considered uplands, and thus potentially contain Plestiodon reynoldsi (Sand Skinks). Vegetation types were classified using the Florida Land Use, Cover, and Forms Classification System (flDOT1999), and then categorized below to distinguish the major characteristics of each. Open (e.g., sandy) patches are categorized by size (absent, small, large), while all other variables are categorized by relative abundance (absent, present, dominant). All vegetation types were sampled according to their availability at our study site (see text). Vegetation type Canopy Shrub layer Groundcover Open patches Improved pasture Absent Absent Dominant Small Unimproved pasture Absent Absent Dominant Small Palmetto prairie Absent Dominant Dominant Absent/small Shrub and brush Absent Dominant Dominant Absent Pine flatwoods Present Present Present Absent Xeric scrub Absent Dominant Present Large Sand live oak Dominant Present Present Large Figure 1. Outline of our study site in Florida showing the location and shape of our 9 sampling plots and their associated vegetation (A) or soil (B) types. The large white (open) areas represent wetlands, which are not suitable as habitat for Plestiodon reynoldsi (Sand Skink). Note that sampling plots may each contain more than one patch of each vegetation or soil type (e.g., vegetation type was independent of the soil underneath). 2007 D.A. Pike, K.S. Peterman, and J.H. Exum 719 (Fig. 1). Each plot contained plywood coverboards (60 x 60 x 1.25 cm) spread evenly in rows at a density of 64 boards per ha (i.e., 28 m apart). We cleared all groundcover underneath each board so that only a sandy substrate remained, allowing greater visibility of Sand Skink trails. Since vegetation can mask trails, this was especially important in areas with dense herbaceous vegetation where there are few areas of open sand. We sampled coverboards once weekly for four consecutive weeks during the mating season (March–May), when activity is thought to be greatest (Andrews 1994, Sutton et al. 1999, but see Ashton and Telford 2006 for high levels of activity during other months). When sampling, we looked underneath each coverboard to determine whether Sand Skink trails were present. All fieldworkers were trained in identifying trails, and in the rare case that a question arose about a particular trail, two or more researchers reached consensus as to whether the trail was made by a Sand Skink. At each trail, we recorded the exact location of the coverboard by using a Global Positioning System (Trimble XL GPS; accurate to less than 1 m). We assumed only one skink was responsible for trails underneath each coverboard, resulting in conservative estimates of relative density (Sutton et al. 1999; D.A. Pike, unpubl. data). Additionally, each coverboard was counted only once when calculating measures of relative density to help avoid pseudosampling. Relative density is defined herein as the proportion of unique coverboards with Sand Skink trails over the 4-week sample period. Although we sampled different portions of the site in successive years (2004 or 2005), we were interested in presence and relative density only, so we combined years for analysis. Statistical analysis We spatially mapped all sampling plots and skink locations in ArcView 3.2 using GPS locations. We layered these points onto grounddetermined vegetation types (see Table 1 for vegetation types consistent with the Florida Land Use, Cover, and Forms Classification System; FL DOT1999) and soil types mapped by the Natural Resources Conservation Service (NRCS; see Fig. 2 for a list of all soil types sampled). Each of our 9 sampling plots contained multiple vegetation and soil types (i.e., multiple patches). For statistical purposes, we calculated presence or relative density separately for each vegetation or soil patch within each sampling plot, and treated patches of the same habitat type as replicates (Fig. 1). Since vegetation types were not necessarily related to the underlying soil, numbers of patches and their sizes varied between these two variables. To test the hypothesis that relative density differed among habitat types, taking into account the size of each habitat patch, we used analysis of covariance (ANCOVA). Habitat type (for vegetation and soil) was the factor, with the size of each habitat patch as the covariate and relative density of Sand Skinks as the dependent variable. We also determined the 720 Southeastern Naturalist Vol. 6, No. 4 proportion of each habitat type sampled and the availability of each using ArcView, and tested whether we surveyed habitats according to their availability at our site using chi-squre goodness of fit tests. We also used chi-squre tests to determine whether the presence of Sand Skinks differed among habitat types (for vegetation and soil). Alpha was set at 0.05 for all tests, and assumptions of statistical tests were checked and met unless otherwise stated. Results Overall survey results We checked 2975 coverboards once weekly for 4 consecutive weeks, resulting in a total of 11,900 possible encounters with Sand Skink trails. During this time we encountered trails underneath 726 individual coverboards (24.5% of possible boards; Table 2). We sampled all vegetation types present on site (n = 7 types) that were designated as uplands using the Florida Land Use, Forms, and Classification System (Table 1; FL DOT 1999). Additionally, we sampled all soil types present at our study site (n = 10; Table 2, Fig. 1). Localized vegetation We found Sand Skinks in 7 out of the 8 sampling plots (87.5%), and presence did not depend upon the vegetation type contained within the sampling plots (χ2 = 8.15, df = 10, P = 0.61; Table 2). Sand Skinks were present in each vegetation type sampled, and those that were not sampled completely were sampled in equal proportion to their availability (χ2 = 6.75, df Figure 2. Soil suitability index for Sand Skinks in Lake County, FL. “Mixed” soil refers to a combination of Pompano, Felda, and Oklawaha soils. Index values were calculated by taking the expected number of plots occupied by skinks (as calculated by a chi-square goodness-of-fit test) and subtracting them from the observed number of sampling plots occupied (after Kuhnz et al. 2005). In soil types with an [observed - expected] value of 0, skinks were found in similar proportions to the calculated expected values. No soil types contained fewer skinks than expected. Note that these values are relative to other soils at our site, and may not represent absolute soil suitability. 2007 D.A. Pike, K.S. Peterman, and J.H. Exum 721 = 3, P = 0.08). Thus, our sampling design controlled for area by sampling representative areas of each available vegetation type. Additionally, Sand Skink relative density did not differ by vegetation type (ANCOVA: F10,30 = 0.74, P = 0.68), but depended upon the amount of area sampled (F1,30 = 5.26, P = 0.03; habitat * area interaction: F9,23 = 0.11, P = 0.99). The larger the area sampled, the higher the relative density, regardless of vegetation type (Fig. 3). A mean relative density of 48.8 ± 9.2 skinks/ha was found within each vegetation type. Underlying soils The proportion of habitat occupied was dependent upon soil type (χ2 = 28.1, df = 10, P = 0.002); however, all soil types contained Sand Skinks (Fig. 2) except one plot containing hydric soils (a mixture of Pompano, Felda, and Oklawaha depressional soils, and Myakka sands; Table 2). All of the occupied soil types were xeric (dry). Most soil types contained more-occupied habitat than would be expected based on values calculated from chi-square tests (Fig. 2). Because a chi-squre test compares the actual proportion of occupied habitat to that expected based on the amount and type surveyed, this indicates that each soil type containing Sand Skinks (n = 8; Table 2, Fig. 2) contained greater proportions of occupied habitat than would be expected by random chance. Sand Skink relative density did not differ by soil type (ANCOVA: F8,13 = 1.00, P = 0.48), but Table 2. Results of our surveys for Plestiodon reynoldsi (Sand Skink) in Florida showing the amount of area sampled in each vegetation and soil type, the number of coverboards sampled in each habitat, and the number and proportion of coverboards under which we found P. reynoldsi or signs of their presence. See text for details regarding the size and density of coverboards and sizes of the habitat patches that were sampled. Coverboards Coverboards showing Habitat Ha sampled (%) sampled P. reynoldsi (%) Vegetation Improved pasture 21.65 (73.1) 1386 523 (37.7) Shrub and brush 0.94 (8.7) 60 4 (6.6) Palmetto prairie 2.13 (40.0) 137 19 (13.9) Pine flatwoods 5.25 (54.8) 336 10 (3.0) Sand live oak 9.07 (99.6) 580 117 (20.2) Unimproved pasture 2.78 (27.9) 298 40 (22.4) Xeric scrub 4.66 (99.8) 178 18 (6.0) Soils Anclote 2.82 (5.8) 180 55 (30.5) Candler 12.26 (79.9) 784 263 (33.5) Immokalee 7.14 (22.0) 457 16 (3.5) Lake 2.50 (99.1) 160 99 (61.9) Myakka 0.70 (17.1) 49 0 (0.0) Placid and Myakka 0.16 (3.1) 10 3 (29.7) Pompano 4.16 (88.3) 266 84 (31.6) Pompano, Felda, and Oklawaha 0.14 (2.3) 9 0 (0.0) Tavares 13.87 (48.2) 888 188 (21.2) 722 Southeastern Naturalist Vol. 6, No. 4 depended upon the amount of area sampled (F1,13 = 19.54, P < 0.001; soil * area interaction: F3,11 = 0.54, P = 0.66). The larger the area sampled, the higher the relative abundance of Sand Skinks, regardless of the specific type of xeric soil (Fig. 3). Discussion Our results support previous studies demonstrating that habitatspecific endemic species can persist in human-altered habitats (Franken and Hik 2004, Scott et al. 2006). Specifically, the Sand Skink was found in all of the human-altered vegetation types we surveyed, which are representative of disturbed lands within central Florida. This result provides evidence that over small spatial scales the distribution of Sand Skinks may not be limited solely to specific vegetation types or ecotonal boundaries, and may be more general than previously thought. This confirms and strengthens the hypothesis of Campbell and Christman (1982), who suspected that soil type may be more important than vegetation in determining Sand Skink distribution, but limited their investigations to natural habitats in different stages of disturbance-induced succession. Although previous studies found high abundances in natural habitats (e.g., McCoy et al. 1999), relative densities did not differ among any of our altered vegetation or soil types, suggesting that there is no difference in perceived quality among these habitats. The behavioral ecology of fossorial reptiles may partially explain our observed patterns. For example, the space use of fossorial reptiles may Figure 3. Relationship between the size of the area sampled and Sand Skink relative density for habitats characterized by vegetation (triangles, solid line) and soil types (squares, dotted line). See text for statistical results. 2007 D.A. Pike, K.S. Peterman, and J.H. Exum 723 be extremely limited; their lives are spent underneath the soil surface in relatively small habitat patches (Andrews 1994, Branch et al. 2003, Campbell and Christman 1982). Therefore, soil composition may be the most important factor limiting fossorial lizard distributions, rather than the composition of the plant communities growing on the sandy soils (Campbell and Christman 1982, Lee 1969). This ecological characteristic may allow persistence of Sand Skink populations in human-altered vegetation types well after alteration occurs, especially when underlain by xeric soils. Since Sand Skinks were found in not only all of the xeric soil types sampled, but in all vegetation types as well, we cannot elucidate whether one of these factors is more important in shaping local-scale distribution. Environmental impact surveys searching for protected species prior to extensive habitat manipulation (e.g., land clearing, development) must take into consideration the fact that endemic species may also exist within severely modified habitat types, even if they are not considered optimal. Environmental assessments have the obligation to fairly determine the presence and distribution of focal species, usually species protected by law. However, when protected species have the reputation of being habitat- specific (often determined by vegetation type), researchers may focus on intact natural habitats, and ignore adjacent areas. Our data revealed that this approach to environmental impact surveys would underestimate the presence of Sand Skinks, and potentially other sympatric endemics, within the landscape. Rapid destruction of habitat is leaving few areas of intact natural habitat available for conservation purposes (Myers and White 1987, Shine et al. 1998). However, the finding that an endemic, and supposedly habitatspecific, lizard species persists in human-altered landscapes suggests that habitat protection and restoration of areas including cattle ranchland and citrus groves is potentially of considerable conservation value. This may be critical in extreme cases where limited tracts of intact natural lands are available for purchase by conservation organizations or public entities for conservation. In fact, our recent efforts to find natural habitat appropriate for conserving populations of Sand Skink indicate that this is the case in Florida, and thus, habitat restoration in altered lands containing populations, if successful, may be the most pressing conservation measure for this and ecologically similar species. In addition to protected species, many other declining or even common species composing portions of communities are present in altered habitats (e.g., Ballinger and Watts 1995, Fitch 2006). Therefore, protection and restoration efforts may also benefit the composition of rare ecological communities, in addition to single focal species. With a thorough understanding of rare species distributions, habitat associations at multiple scales, and ecological tolerances, the probability 724 Southeastern Naturalist Vol. 6, No. 4 that conservation practices will succeed increases (Pringle et al. 2003, Webb et al. 2005). Additionally, the more habitat types (especially human-altered habitats) that contain rare species, the higher the chances are that habitat can be obtained, protected, and restored in such a way as to ensure persistence. We emphasize that natural habitats should be the focus of conservation efforts whenever possible; however, human-altered lands should not be written off as unsuitable for rare species before thorough research indicates otherwise. Acknowledgments Access to land and funding were provided by Ms. Virginia Dio and Ms. Liz Tapado. K. Nelson, R. Mejeur, and K. Muddle were responsible for much of the fieldwork; H.R. Mushinsky provided guidance and insight during our research. We are especially grateful to K.G. Ashton and E.A. Roznik for greatly improving the clarity and quality of a previous draft of this manuscript. Monkey Bar Orlando provided a work environment conducive to the generation of novel ideas, along with an excellent writing atmosphere. Literature Cited Andrews, R.M. 1994. Activity and thermal biology of the Sand-swimming Skink Neoseps reynoldsi: Diel and seasonal patterns. Copeia 1994:91–99. Ashton, K.G. 2005. Life history of a fossorial lizard, Neoseps reynoldsi. Journal of Herpetology 39:389–395. Ashton, K.G., and S.R. Telford, Jr. 2006. Monthly and daily activity of a fossorial lizard, Neoseps reynoldsi. Southeastern Naturalist 5:175–183. Ballinger, R.E., and K.S. Watts. 1995. Path to extinction: Impact of vegetational change on lizard populations on Arapaho Prarie in the Nebraska sandhills. American Midland Naturalist 134:413–417. Branch, L.C., A.-M. Clark, P.E. Moler, and B.W. Bowen. 2003. Fragmented landscapes, habitat specificity, and conservation genetics of three lizards in Florida scrub. Conservation Genetics 4:199–212. Campbell, H.W., and S.P. Christman. 1982. The herpetological components of Florida sandhill and sand pine scrub associations. Pp. 163–171, In N.J. Scott, Jr. (Ed.). Herpetological Communities. US Fish and Wildlife Service. Washington, DC. Cooper, B.W. 1953. Notes on the life history of the lizard, Neoseps reynoldsi Stejneger. Quarterly Journal of the Florida Academy of Sciences 16:235–238. Fitch, H.S. 2006. Collapse of a fauna: Reptiles and turtles of the University of Kansas Natural History Reservation. Journal of Kansas Herpetology 17:10– 13. Florida Department of Transportation (flDOT). 1999. Florida Land Use, Cover, and Forms Classification System. Surveying and Mapping Office, Tallahassee, FL. Franken, R.J., and D.S. Hik. 2004. Influence of habitat quality, patch size, and connectivity on colonization and extinction dynamics of collared pikas Ochotona collaris. Journal of Animal Ecology 73:889–896. 2007 D.A. Pike, K.S. Peterman, and J.H. Exum 725 Gibbons, J.W., and P.W. Stangel (Eds.). 1999. Conserving amphibians and reptiles in the new millennium. In Proceedings of the Partners in Amphibian and Reptile Conservation (PARC) Conference, 2–4 June 1999, Atlanta, GA. Savannah River Ecology Laboratory, Herp Outreach Publication 2, Aiken, SC. Jäggi, C., and B. Bauer. 1999. Overgrowing forest as a possible cause for the local extinction of Vipera aspis in the northern Swiss Jura mountains. Amphibia- Reptilia 20:25–34. Kuhnz, L.A., R.K. Burton, P.N. Slattery, and J.M. Oakden. 2005. Microhabitats and population densities of California legless lizards, with comments on effectiveness of various techniques for estimating numbers of fossorial reptiles. Journal of Herpetology 39:395–402. Lee, D.S. 1969. Moisture toleration: A possible key to dispersal ability in three fossorial lizards. Bulletin of the Maryland Herpetological Society 5:53–56. McCoy, E.D., P.E. Sutton, and H.R. Mushinsky. 1999. The role of guesswork in conserving the threatened Sand Skink. Conservation Biology 13:190–194. Means, D.B., and D. Simberloff. 1987. The peninsula effect: Habitat-correlated species decline in Florida’s herpetofauna. Journal of Biogeography 14:551–568. Myers, R.L., and D.L. White. 1987. Landscape history and changes in sandhill vegetation in north-central and south-central Florida. Bulletin of the Torrey Botanical Club 114:21–32. Panzer, R., D. Stillwaugh, R. Gnaedinger, and G. Derkovitz. 1995. Prevalence of remnant dependence among the prairie-inhabiting and savanna-inhabiting insects of the Chicago region. Natural Areas Journal 15:101–116. Pringle, R.M., J.K. Webb, and R. Shine. 2003. Canopy structure, microclimate, and habitat selection by a nocturnal snake, Hoplocephalus bungaroides. Ecology 84:2668–2679. Ries, L., D.M. Debinski, and M.L. Wieland. 2001. Conservation value of roadside prairie restoration to butterfly communities. Conservation Biology 15:401–411. Schmitz, A., P. Mausfeld, and D. Embert. 2004. Molecular studies on the genus Eumeces Wiegmann, 1834: Phylogenetic relationships and taxonomic implications. Hamadryad 28:73–89. Schneider, M.F. 2001. Habitat loss, fragmentation, and predator impact: Spatial implications for prey conservation. Journal of Applied Ecology 38:720–735. Scott, D.M., D. Brown, S. Mahood, B. Denton, A. Silburn, and F. Rakotondraparany. 2006. The impacts of forest clearance on lizard, small mammal, and bird communities in the arid spiny forest, southern Madagascar. Biological Conservation 127:72–87. Shine, R., J.K. Webb, M. Fitzgerald, and J. Sumner. 1998. The impact of bush-rock removal on an endangered snake species, Hoplocephalus bungaroides. Wildlife Research 25:285–295. Sutton, P.E., H.R. Mushinsky, and E.D. McCoy. 1999. Comparing the use of pitfall drift fences and cover boards for sampling the threatened Sand Skink (Neoseps reynoldsi). Herpetological Review 30:149–151. Telford, S.R. Jr. 1959. A study of the Sand Skink, Neoseps reynoldsi Stejneger. Copeia 1959:110–119. Telford, S.R. Jr. 1962. New locality records for the Sand Skink (Neoseps reynoldsi) in central Florida, with comments on the habitat. Quarterly Journal of the Florida Academy of Sciences 25:76–77. 726 Southeastern Naturalist Vol. 6, No. 4 United States Fish and Wildlife Service. 1993. Recovery Plan for the Sand Skink and the Blue-tailed Mole Skink. US Fish and Wildlife Service, Atlanta, GA. Webb, J.K., and R. Shine. 1998. Using thermal ecology to predict retreat-site selection by an endangered snake species. Biological Conservation 86:233–242. Webb, J.K., R. Shine, and R.M. Pringle. 2005. Canopy removal restores habitat quality for an endangered snake in a fire-suppressed landscape. Copeia 2005:893–899.