2010 NORTHEASTERN NATURALIST 17(4):629–638 Salamander Diversity at C.F. Phelps Wildlife Management Area, Fauquier and Culpeper Counties, Virginia Jay D. McGhee1,* and Michael D. Killian2 Abstract - Salamander guilds are important components of ecosystems and may be declining in Virginia. Consequently, information on salamander diversity and abundance is needed. Our objective was to assess salamander diversity at one site in the Rappahannock River watershed: the C.F. Phelps Wildlife Management Area. We randomly selected stream and upland terrestrial sites to run 50-m transects, for both quadrat and natural cover searches. We assessed diversity using a Shannon-Weiner index on all captures (larval and adults) and assessed diversity in on-site catchments. We found 11 of 13 expected species, with Ĥ ' = 1.33 ± 0.05 SD, Ĵ ' = 0.55 for all captures, and Ĥ ' = 1.18 ± 0.08 SD, Ĵ ' = 0.49 for non-larval diversity. A single catchment (Fishing Run) was considered more diverse than other catchments on site. We conclude that C.F. Phelps Wildlife Management Area supports a relatively diverse salamander community. Management efforts should focus on maintaining stream structural diversity and monitoring the effects of agricultural activities such as fertilizer use, erosion, and habitat fragmentation and loss. Introduction Salamander communities often represent an important component within ecosystems and function as substantial links in food chains, maintain lower-level prey diversity, slow leaf-litter processing, affect soil dynamics, and act as pathways for energy transfer between aquatic and terrestrial habitats (Davic and Welsh 2004). Despite their importance, however, we have an incomplete understanding of salamander life histories and community interactions. This is particularly problematic in the face of global amphibian declines, the solutions for which require a sophisticated comprehension of salamander diversity, ecology, and monitoring methods (Heyer and Murphy 2005, Williams and Berkson 2004). Intensive surveys of potential habitat are therefore beneficial to increase our understanding of current levels of salamander diversity, which we define as the combination of species richness (the number of species) and evenness (the relative abundance of species), in a community (Lloyd and Ghelardi 1964, McIntosh 1967). In the Rappahannock River watershed of northern Virginia, short-term surveys of amphibian diversity have been conducted at military bases out of concern for possible declines, but no results of extended surveys have been published (Mitchell 1998). Our overall goal was to conduct 1Randolph-Macon College, Ashland, VA 23005. 2Department of Biological Sciences, University of Mary Washington, Fredericksburg, VA 22401. *Corresponding author - jaymcghee@rmc.edu. 630 Northeastern Naturalist Vol. 17, No. 4 a more intensive two-year survey to assess salamander diversity on a wildlife management area within this region. Assessing species diversity can be expensive and time-consuming, so it is important to establish the sampling intensity required to adequately estimate this parameter. In addition, it is desirable to assess diversity on multiple habitat scales. This approach can enable researchers to determine whether species diversity is primarily a product of richness within sites (alpha diversity) or species turnover between sites (beta diversity) and allow managers to pinpoint those sites that require the most protection (Gering et al. 2003). Our study had three objectives: 1) to determine salamander species diversity at C.F. Phelps Wildlife Management Area (WMA) in the Piedmont portion of the Rappahannock River watershed, 2) to determine differences in salamander diversity within the study site in order to identify areas of management interest, and 3) to assess the number of samples required to detect the total number of salamander species (species richness) and species diversity (species richness and evenness) detected by the survey. Field Site Description Seventeen salamander species (approximately one-third of all species known from Virginia) have been recorded within the Rappahannock River drainage, located largely within the Piedmont region of the state (Mitchell and Reay 1999). The C.F. Phelps WMA is located within this drainage and borders the river. This WMA of 1837 ha (4539 acres) is located in southern Fauquier and eastern Culpeper counties. Several small streams cross the property, and empty into the Rappahannock River, which borders the WMA western edge. The land consists of pine and hardwood forest of mixed age, approximately 400 ha of open habitat formerly used for agriculture, and a 1.2-ha man-made pond situated in the middle of the WMA. Elevations range from 200–400 m above sea level. Thirteen species of salamanders are expected to occur on the property, based on past species occurrences in Fauquier and surrounding counties (Mitchell and Reay 1999): Ambystoma maculatum Shaw (Spotted Salamander), A. opacum Gravenhorst (Marbled Salamander), Desmognathus fuscus Rafinesque (Northern Dusky Salamander), D. monticola Dunn (Seal Salamander), Eurycea bislineata Green (Northern Two-lined Salamander), E. guttolineata Holbrook (Three-lined Salamander), Gyrinophilus porphyriticus Green (Spring Salamander), Hemidactylium scutatum Temminck and Schlegel (Four-toed Salamander), Notophthalmus viridescens Rafinesque (Red-Spotted Newt), Plethodon cinereus Green (Red-backed Salamander), P. cylindraceus Harlen (White-spotted Slimy Salamander), Pseudotriton montanus Baird 1850 (Mud Salamander), and P. ruber Sonnini de Manoncourt and Latreille (Red Salamander). 2010 J.D. McGhee and M.D. Killian 631 Methods Salamanders were sampled from an aggregate of stream and terrestrial transects. We selected sampling locations by first locating stream sites occurring nearest to a randomly selected GPS location on the property and then moving up or downstream by a randomly selected distance of 0–50 m. From these sites we ran a 50-m transect downstream, and sampled five 1-m2 quadrats, laid randomly within 10-m increments (Jaeger 1994, Jaeger and Inger 1994, Mitchell 2000). Quadrats were placed such that 1/2 to 1/3 of their area covered the bank bordering the stream and, depending on stream width, all together they covered the entire stream or covered only one side of the stream. We searched quadrats by removing large cover objects such as rocks and decaying wood, searching through leaf pack, and dragging fine-mesh aquatic dip nets across the stream bottom (Mitchell 2000). In August 2007, we realized that stream banks might provide adequate habitat for some species and decided to incorporate another 1-m2 quadrat search on the banks bordering each stream quadrat. This addition essentially increased our quadrat samples to 10 per stream, five in the stream and/or bordering it and five on the bank of the stream. We searched these bank quadrats by moving large cover objects such as rocks and decaying wood and searching through leaf litter. We selected terrestrial transect sites by randomly selecting an azimuth and meter distance (0–200 m) from the starting point of the stream site. We employed two types of terrestrial survey methods: leaf-litter quadrat searches (LLQ) and natural-cover searches (NC). For LLQ quadrat searches we laid a 50-m transect and sampled five 1-m2 quadrats, laid randomly within 10-m increments (Jaeger 1994, Jaeger and Inger 1994, Mitchell 2000). Fifty-meter natural-cover searches were conducted 10 m distant and parallel to quadrat searches. Observers searched a 3-m-wide strip by overturning all natural-cover objects, such as large rocks, pieces of bark, and fallen limbs and logs (Heatwole 1962, Hyde and Simons 2001). In addition, we searched opportunistically on cool, wet nights. Night-search transect locations were randomly selected but constrained to occur within 50 m of a road on the WMA. We visually searched the ground and under natural-cover objects within a 50-m transect covering a 3-m width (Hyde and Simons 2001). For all transect types, we identified captured salamanders to species, and measured snout–vent and total lengths (Petranka 1998). We analyzed capture data using the Shannon-Weiner estimate (Ĥ '), which measures both species richness and evenness (Krebs 1999, Lloyd et al. 1968). The Shannon-Weiner estimate derives from information theory, which measures the amount of order in a system or, specific to biodiversity, the variability of species in a system (Krebs 1999, Margalef 1958). The estimate increases with both the number of species captured—an estimate of species richness—and with the degree to which individuals captured are 632 Northeastern Naturalist Vol. 17, No. 4 evenly distributed across all detected species (Krebs 1999). The Ĥ ' can reach large values, but can be standardized to range between 0–1 by dividing by the maximum possible Ĥ ' for all species captured in a study. This value of relative diversity (Ĥ '/Ĥ max) we refer to as Ĵ ' (Zar 1999). To determine differences in diversity within the study area, we divided it into five stream catchments, from largest to smallest: Persimmon Run, Fishing Run, Mine Run, Western Tributaries (a set of small unnamed streams on the western side of the property), and Eastern Stream. We counted the number of species found in each area and calculated Shannon- Weiner indices for these smaller catchments. For the two catchments that appeared to carry the most species, we used a Monte Carlo rarefaction analysis to determine the most diverse habitat, presumably the habitat of greatest management concern (Hurlbert 1971, Tipper 1979). For the watershed with the largest number of individuals, we used ECOSIM to randomly sample from the known capture data, and restricted our number of captures to the lower total sample size of the catchment with which we wished to make a comparison (catchment with the second-highest number of captures) (Gotelli and Entsminger 2004). We performed this random sampling for 1000 iterations, and calculated 1000 Shannon-Weiner estimates. We then calculated a 95% confidence interval for the estimate and compared it to the catchment with the second-highest number of captures. If the smaller community falls within the 95% CI, this implies no statistical difference between the sites (Gotelli and Entsminger 2004). The advantage of this method is that it makes no a priori assumptions regarding the probability distribution of the data (i.e., Gaussian distribution), a difficult assumption to justify when the number of species (data points) is relatively small. Finally, we graphically analyzed the number of transects required to capture all the species found during the study, using the Shannon-Weiner index as our dependent variable (Krebs 1999, Scott 1994). Results We suveyed from 13 April 2007 to 21 April 2009. We sampled 78 stream transects with 390 stream quadrats and 295 stream-bank quadrats. We sampled 88 LLQ transects with 440 quadrats, 89 NC transects, and 39 night transects. We captured a total of 516 salamanders, comprised of 129 adults and 387 subadults (juveniles and larvae). We found 11 of the 13 species expected based on Mitchell and Reay (1999). We did not capture any Spring or Mud Salamanders. Red-backed Salamander dominated our captures (Fig. 1). We calculated a Ĥ ' = 1.33 ± 0.05 SD. We estimated relative diversity (Ĵ ') as 0.55, meaning our distribution of captured species was 55% of the maximum (most uniform) distribution possible. Within our samples, we noticed a tendency for Two-lined Salamander and Spotted Salamander larvae to occur in a clumped distribution, possibly after hatching from single egg masses. This tendency for members of the same species to occur together acts to 2010 J.D. McGhee and M.D. Killian 633 negatively bias Ĥ ' (Krebs 1999). To offset this, we decided to calculate a Shannon-Weiner estimate after removing larvae of all species (Ĥ ' = 1.18 ± 0.08 SD, Ĵ ' = 0.49 ). Smaller catchments within the property differed in both number of species and non-larval individuals captured (species, individuals, Ĥ ' ± SD, n = number of transects): Fishing Run (8, 53, 1.40 ± 0.13, 85), Eastern Stream (5, 27, 1.16 ± 0.15, 23), Mine Run (5, 32, 0.88 ± 0.19, 18), Persimmon Run (6, 121, 0.84 ± 0.10, 93), and the Western Tributaries (4, 22, 0.75 ± 0.20, 34). We compared Persimmon Run to Fishing Run by randomly sampling capture data from Persimmon Run, limited to a total capture of 53 individuals. From 1000 iterations, we calculated Ĥ ' = 0.81 ± 0.11 SD. This resulted in confidence intervals ranging from 0.56–1.02, implying significantly less diversity in Persimmon Run than in Fishing Run. We detected only one species unique to a particular watershed: the Four-toed Salamander, captured in the Eastern Stream catchment. We captured all species found on site with 94 transects; however, 10 of these species were captured within only 30 transects, and nine within 13 transects. Our Shannon-Weiner estimates ranged from between 0.67–1.78 over the entire sampling period. This range decreases to 1.33–1.45 after all observed species were captured, but does not indicate that the estimate has reached its asymptote (Fig. 2). Discussion Discounting larvae, the distribution of individuals among the species we detected was 49% of the maximum distribution possible, and skewed Figure 1. The number of individuals, excluding larvae, captured per species on the C.F. Phelps Wildlife Management Area, Fauquier and Culpeper counties, VA, April 2007–April 2009. 634 Northeastern Naturalist Vol. 17, No. 4 heavily toward the Red-backed Salamander. This pattern of a single numerically dominant species combined with many rare species is well documented (MacArthur 1972, Preston 1948) and matches the salamander- guild patterns described by Davic and Welsh (2004). Based on their review of five studies with at least 1000 captures in forested landscapes, they found that a single species—often the Red-backed Salamander— tended to dominate salamander communities. Those factors that determine shifts in numerical dominance among salamander species are largely unknown (Davic and Welsh 2004). Terrestrial salamanders are thought to prefer areas with many cover objects, high soil moisture, neutral soil pH, lower temperatures, and ready access to lower soil layers as predation refugia (Bogert 1952, Heatwole 1962, Mathewson 2009, Spotila 1972). The Red-backed Salamander is smaller than many of our expected terrestrial species, is a habitat generalist, and so may better exploit soil systems with these features than larger species in the area, particularly if prey or some other habitat feature is limited. Alternatively, McGill and Collins (2003) suggest that this pattern can be explained based on a combination of the independent distribution of species across a landscape and the pattern of species ranges: high density concentrations of individuals diffusing outward to lower density concentrations. This explanation would mean individuals of a single species are common in a few habitats, and rarer in many others. Given these assumptions, sampling at any particular site will result in collecting many individuals from the small set of species common for that area and collecting a few individuals from each of a wider range of rarer species, resulting in the hollow curve pattern we obtained. Figure 2. Shannon-Weiner estimates as sampling effort increases on the C.F. Phelps Wildlife Management Area, Fauquier and Culpeper counties, VA, April 2007–April 2009. The cumulative number of captured species are shown with arrows. 2010 J.D. McGhee and M.D. Killian 635 Our survey failed to detect Spring and Mud Salamanders. C.F. Phelps WMA occurs on the extreme eastern edge of the range of the Spring Salamander in Virginia, and this species is likely to occur only rarely in the area (VDGIF 2010). Similarly, C.F. Phelps WMA lies just outside the western edge of the range of Pseudotriton m. montanus (Eastern Mud Salamander) in Virginia, and our study area had few of the bog, swamp, or floodplain forest habitats associated with this sub-species (Petranka 1998, VDGIF 2010). Consequently, we anticipated this sub-species would be rare in the area. We were able to detect 84% of the species we expected to find, however, which implies that diversity patterns have not changed substantially since the Mitchell and Reay (1999) atlas was published. We captured fewer adult Marbled and Spotted Salamanders than we had initially expected. Adults of both species tend to occur in bottomland forests and floodplains, but are most active on cool rainy nights, and only seasonally, which may account for their relatively small numbers in our total catches (Petranka 1998). Our results agree loosely with the Mitchell (1998) survey of three military sites near our study area. He reported a diverse amphibian complex with no evidence of aquatic amphibian declines at Quantico Marine Corps Base, approximately 40 km distant from our study site, and Fort A.P. Hill and Fort Belvoir, each approximately 60 km distant. However, relative abundances differed between sites, with the Mitchell aquatic captures being dominated by Spotted, Seal, Northern Dusky, Two-lined, and Spring Salamanders, whereas our aquatic samples were dominated by the Two-lined Salamander. This variability would be expected to the extent that habitat differs between sites, but Mitchell performed a set of short-term surveys over a wider geographic area, and the difference in methodologies makes it difficult to compare diversity between these sites. Although the general pattern of species dominance remained the same, we found that stream catchments on our study area differed in salamander diversity. The largest streams, Persimmon Run and Fishing Run, were the most diverse, and Fishing Run was significantly more diverse than Persimmon Run. Fishing Run sites were often shallower than those in Persimmon Run and had a greater variation in habitat, including seeps and Castor canadensis Kuhl (Beaver)-modified and deeper waters. We suggest that stream structural diversity be maintained at Fishing Run, and action be taken to minimize agricultural and roadway runoff to this stream. Ephemeral streams should be maintained for Marbled Salamander, along with clear, shallow streams with seepages and bog habitats for Red and Three-lined Salamanders (Petranka 1998). Coarse woody debris and fallen logs will provide cover objects for terrestrial species such as the Red-backed and Slimy Salamanders (Petranka 1998). The majority of species were found in multiple or all catchments, implying that alpha, or within catchment, diversity explains more of our diversity estimate, than beta, or between catchment, diversity. We did find one species 636 Northeastern Naturalist Vol. 17, No. 4 unique to a catchment: the Four-toed Salamander on the Eastern Stream site at the southern edge of the property. Although the status of the Four-toed Salamander is unknown, its specialized use of moss mats, particularly of Sphagnum moss near still water, for egg-laying, means populations are potentially at risk from habitat loss or degradation (Blanchard 1923). Petranka (1998) suggests managers should retain a mature deciduous tree canopy over potential habitat to retain moisture, allow the deposition of coarse woody debris, and promote moss growth by raising earth to form hummocks around vernal pools and marshes. Sampling effort is an important component of survey design. We found with our methods that the majority of species could be captured within 13 or fewer transects, while rarer species or those most difficult to detect with our sampling design, required considerably greater effort. A precise estimate of diversity, to include a precise estimate of evenness, would be almost impossible to achieve in a short-term survey. While species richness estimates may be achieved quickly, this will tend to positively bias the Shannon-Weiner estimate because species numbers will be increased to a greater extent than the relative abundances of those species. We suggest that, at a minimum, long-term intensive surveys are important to provide validation data for future common short-term and less intensive surveys. Acknowledgments We thank Joe Ferdinandsen, manager of C.F. Phelps WMA, for his cooperation, along with R. Hughes. We thank University of Mary Washington students Carly Byers, Sarah Almahdali, Jennifer Clary, Hillary Adams, and Ramsey Hanna for their field work assistance. We thank the students of J.D. McGhee’s animal ecology classes in Fall 2007 and 2008 for their help in field sampling. Three anonymous reviewers made numerous and useful comments during manuscript review. Literature Cited Blanchard, F.N. 1923. The life history of the Four-toed Salamander. American Naturalist 57:262–268. Bogert, C.M. 1952. Relative abundance, habitats, and normal thermal levels of some Virginian salamanders. Ecology 33:16–30. Davic, R.D., and H.H. Welsh, Jr. 2004. On the ecological roles of salamanders. Annual Review of Ecology, Evolution, and Systematics 35:405–434. Gering, J.C., T.O. Crist, and J.A. Veech. 2003. Additive partitioning of species diversity across multiple spatial scales: Implications for regional conservation of biodiversity. Conservation Biology 17:488–499. Gotelli, N.J., and G.L. Entsminger. 2004. EcoSim: Null models software for ecology. Version 7. Acquired Intelligence, Inc. and Kesey-Bear, Jericho, VT 05465. Available online at http://garyentsminger.com/ecosim/index.htm. Accessed 11 September 2009. Heatwole, H. 1962. Environmental factors influencing local distribution and activity of the salamander, Plethodon cinereus. Ecology 43:460–472. Heyer, W.R., and J.B. Murphy. 2005. Declining amphibian populations task force. Pp. 17–21, In M. Lannoo (Ed.). Amphibian Declines: The Conservation Status of United States Species. University of California Press, Berkeley, CA. 2010 J.D. McGhee and M.D. Killian 637 Hurlbert, S.H. 1971. The nonconcept of species diversity: A critique and alternative parameters. Ecology 52:577–586. Hyde, E.J., and T.R. Simons. 2001. Sampling plethodontid salamanders: Sources of variability. Journal of Wildlife Management 65:624–632. Jaeger, R.G. 1994. Transect sampling. Pp. 103–106, In W.R. Heyer, M.A. Donnelly, R.W. McDiarmid, L.C. Hayek, and M.S. Foster (Eds.). Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Smithsonian Institution Press, Washington, DC. Jaeger, R.G., and R.F. Inger. 1994. Quadrat sampling. Pp. 97–102, In W.R. Heyer, M.A. Donnelly, R.W. McDiarmid, L.C. Hayek, and M.S. Foster (Eds.). Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Smithsonian Institution Press, Washington, DC. Krebs, C.J. 1999. Ecological Methodology, 2nd Edition. Addison-Wesley, Menlo Park, CA. Lloyd, M., and R.J. Ghelardi. 1964. A table for calculating the “equitability” component of species diversity. Journal of Animal Ecology 33:217–225. Lloyd, M., J.H. Zar, and J.R. Karr. 1968. On the calculation of information-theoretical measures of diversity. American Midland Naturalist 79:257–272. MacArthur, R.H. 1972. Geographical Ecology: Patterns in the Distribution of Species. Princeton University Press, Princeton, NJ. Margalef, D.R. 1958. Information theory in ecology. General Systems 3:36–71. Mathewson, B. 2009. The relative abundance of Eastern Red-backed Salamanders in Eastern Hemlock-dominated and mixed deciduous forests at Harvard forest. Northeastern Naturalist 16:1–12. McGill, B., and C. Collins. 2003. A unified theory for macroecology based on spatial patterns of abundance. Evolutionary Ecology Research 5:469–492. McIntosh, R.P. 1967. An index of diversity and the relation of certain concepts to diversity. Ecology 48:392–404. Mitchell, J.C. 1998. Amphibian decline in the Mid-Atlantic Region: monitoring and management of a sensitive resource. Final Report, Legacy Resource Management Program, US Department of Defense, Arlington, VA. Mitchell, J.C. 2000. Amphibian Monitoring Methods and Field Guide. Smithsonian National Zoological Park Conservation and Research Center, Front Royal, VA. Mitchell, J.C., and K.K. Reay. 1999. Atlas of Amphibians and Reptiles in Virginia. Special Publication Number 1, Virginia Department of Game and Inland Fisheries, Richmond, VA. Petranka, J.W. 1998. Salamanders of the United States and Canada. Smithsonian Institution Press, Washington, DC. Preston, F.W. 1948. The commonness, and rarity, of species. Ecology 29:254–283. Scott, N.J., Jr. 1994. Complete species inventories. Pp. 78–84, In W.R. Heyer, M.A. Donnelly, R.W. McDiarmid, L.C. Hayek, and M.S. Foster (Eds.). Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Smithsonian Institution Press, Washington, DC. Spotila, J.R. 1972. Role of temperature and water in the ecology of lungless salamanders. Ecological Monographs 42:95–125. Tipper, J.C. 1979. Rarefaction and rarefiction: The use and abuse of a method in paleoecology. Paleobiology 5:423–434. 638 Northeastern Naturalist Vol. 17, No. 4 Virginia Department of Game and Inland Fisheries (VDGIF). 2010. Eastern Mud Salamander. Available online at http://www.dgif.virginia.gov/wildlife/ information/?s=020069. Accessed 4 May 2010. Williams, A.K., and J. Berkson. 2004. Reducing false absences in survey data: Detection probabilities of Red-backed Salamanders. Journal of Wildlife Management 68:418–428. Zar, J.H. 1999. Biostatistical Analysis, 4th Edition. Prentice Hall, Upper Saddle River, NJ.