Northeastern Naturalist 210 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 22001144 NORTHEASTERN NATURALIST 2V1(o2l). :2211,0 N–2o3. 32 A Regional Study of Diversity and Abundance of Small Mammals in Ohio John D. Harder1,*, Joy K. Kotheimer1, and Ian M. Hamilton1,2 Abstract - The goal of this study was to obtain information on diversity, abundance, and distribution of non-volant small mammals in 4 major habitat types in each of 5 regions of Ohio. We trapped in 31 study areas, representing 39 counties, for 3 consecutive nights for a total of 38,400 trap nights. We established eight 100-m transects (each with 10 live traps, 20 snap traps, and 20 pitfall traps) per study area in woodland, oldfield, grassland-pasture, or restored prairie-wetland habitats. We captured fourteen species of small mammals (shrews and rodents less than 100 g in body mass), but 97% of the 2150 captured consisted of just 4 species: Microtus pennsylvanicus (Meadow Vole; 31%), Peromyscus leucopus (White-footed Mouse; 29%), Blarina brevicauda (Short-tailed Shrew; 21%), and Sorex cinereus (Masked Shrew; 16%). Regional differences in abundance of small mammals (captures/100 trap nights) and species diversity (H') were not significant (P > 0.05). Seven species of interest were captured in low numbers (less than 10) and 2 others, Reithrodontomys humulis (Eastern Harvest Mouse) and Myodes gapperi (Red-backed Vole), were not captured in the course of the 2-year study. Introduction Declining biological diversity is widely recognized as the major challenge in conservation. Concern is often greatest for losses associated with deforestation in developing countries, but nations with well-established conservation programs, such as the United States, are also confronted with this issue. Ecoregion assessments conducted by the United States Forest Service indicate that 29–46% of vertebrate species in the US are at risk for population decline (Manley et al. 2005), and although amphibian declines are widely recognized, evidence warns of declines in mammalian diversity and populations as well (Entwistle and Stephenson 2002). For instance, Schipper et al. (2008) estimated from a global survey that 25% of all mammals are threatened with extinction. Unfortunately, such conclusions are difficult to evaluate because knowledge of the distribution and abundance of mammals in most states and provinces of North America varies from being current and detailed for game and endangered species to incomplete and outdated for nearly all the rest. In recognition of the challenges to conservation presented by habitat loss and global climate change, there is a critical need to accurately assess both spatial and temporal trends in biological diversity in order to develop relevant land-management and conservation policies (Yoccoz et al. 2001). There is, in fact, sustained 1Department of Evolution, Ecology, and Organismal Biology, 2Department of Mathematics, The Ohio State University, 318 W. 12th Avenue, Columbus, OH 43210. *Corresponding author - harder.2@osu.edu. Manuscript Editor: Thomas French Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 211 interest in studies of mammalian diversity not only in remote tropical regions (Bateman et al. 2010, Caro et al. 2001, Ojeda et al. 2003) but also in North America and Europe (McShea et al. 2003, Panzacchi et al. 2010). Long-term studies (Getz et al. 2001, Reed and Slade 2008, Schwartz and Whitson 1987) have advanced our understanding of community ecology and demography, but they are typically geographically limited and focused on a few species. Sampling protocols have been developed for detecting population change for large suites of vertebrate species based on presence-absence monitoring (Manley et al. 2005), but more detailed, taxon-specific surveys are also needed to provide species-abundance data for comparison with results from large-scale surveys. Efforts in mammalian conservation have traditionally focused on large “charismatic” species, even though most mammalian species are small in size. For example, Entwistle and Stephenson (2002) found that only 15% of the papers published in 4 major conservation journals during 1986–1996 featured mammals with a body mass of less than 1 kg. Moreover, current studies of small mammals typically focus on community and population biology and seldom provide regional assessments of diversity and abundance or update systematic collections. For instance, less than 7% of the specimens of the 6 most common species in museum records (bats, insectivores, and small rodents) of Ohio were collected after 1985 (Harder 1997). Unfortunately, the majority of mammalian species lost in the last 400 years (≈75%), as well as those most likely to become extinct in the near future, are small mammals (Entwistle and Stephenson 2002). Thus, the need for systematic, regional or statewide surveys of mammalian diversity in North America is clear. Our study was designed to assess the distribution, abundance, and diversity of 16 species of non-volant small mammals (i.e., shrews and rodents with a body mass = less than 100 g) in 4 major habitat types in each of 5 regions of Ohio. This group includes both common and rare species, but all are susceptible to capture in the array of trap types used in the study. Because the natural communities and land-use patterns of Ohio are similar to those in surrounding states, the findings of this investigation have relevance for the study and conservation of small mammals throughout the region. Field-site Descriptions Physiographic provinces of Ohio Six study areas were established in each of 5 regions in Ohio; these regions represent a natural subdivision of the state because of how they align with the 4 main physiographic provinces in Ohio (Powers 2011) (Fig. 1). The northwest region is primarily low-relief Lake Plains and includes the former Great Black Swamp, the last major area of the state to be drained and cultivated. It is used intensively for corn and soybean production and has the highest proportion of land cover under cultivation (Table 1). Landscape in the northeast region, representing the Glaciated Plateau, is heterogeneous with glacial moraines and abundant wetlands interspersed with pasture, row crops, and large tracts of forest. Climate, as evidenced in temperature and snowfall, is similar to that in northern Pennsylvania and western New York. The southeast region, almost entirely Unglaciated Appalachian Plateau, is the most Northeastern Naturalist 212 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 highly forested region of the state (Table 1), home to the Wayne National Forest with high hills, sandstone cliffs, and winding streams. The southwest region occupies the Till Plains and a small portion of Kentucky’s Bluegrass Region in southern Figure 1. Location of study areas (dots) in each of 5 Regions of Ohio: central (C), northwest (NW), northeast (NE), southeast SE), and southwest (SW) and their relationship to the 4 major physiographic provinces of Ohio (map modified from Powers 2011), and transect layout. Small-mammal trapping was conducted on 8 transects at each of the 31 study areas throughout the State. One Sherman and 2 snap traps were placed at each of 10 stations along each transect. Three 2.5-m drift fences were placed between the central and 3 peripheral pitfall traps, forming arms of an array that were separated by arcs of approximately 120°. Drift fences were omitted on the pitfall trap array placed between stations 5 and 6. Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 213 Adams County. This region features fertile farmland and forested tracts in a gently rolling topography of glacial moraines and a climate similar to that of northern Kentucky. The central region is predominantly Till Plains. Headwaters of the Scioto River and its tributaries originally drained swamp forest in this region, which is now largely devoted to agriculture in areas outside metropolitan Columbus, Ohio’s largest city. The 5 regions also correspond to administrative Districts of the Ohio Division of Wildlife (ODW), which facilitated administrative and logistical support for the project. Study areas We obtained the percent composition of agricultural and natural land-cover types in 30-m x 30-m cells within an 8-km radius from the center of each study area by GIS analysis. We obtained the 2006 National Land Cover Data Set (NLCD) from the United States Department of Agriculture and processed it using the Arc- GIS Spatial Analyst Extension from Environmental Systems Research Institute. We buffered and clipped land-cover data to an 8-km radius that included but did not extend far beyond the margins of the study areas, and summarized data for each study area by count of the cells in the raster dataset using their classification values in the NLCD Classification System. We calculated the data as percentages for each study area. We sampled each study area, approximately 100 km2, with 8 transects (A–H), which we placed within 1 of 4 habitat types recognized in this study: 1) woodland, 2) grassland, 3) oldfield, or 4) restored prairie/wetland. These habitat types were Table 1. Percent cover (mean ± SE) of major natural and agricultural land-use typesA by region of Ohio. Percentages were computed from the cover type in 30-m x 30-m cells within an 8-km radius from the center of each study area and averaged by region. Grassland, and to a lesser extent, cultivated cover types in the NLCD system includes oldfield as recognized in this study. Grassland also includes restored prairie and wetland (wet meadows). Region ForestB GrasslandC CultivatedD WetlandE Central 25.1 ± 7.6 15.6 ± 1.5 59.0 ± 9.6 0.3 ± 0.1 Northwest 12.7 ± 3.5 9.7 ± 3.1 72.0 ± 6.4 5.5 ± 2.7 Northeast 51.3 ± 6.9 17.9 ± 1.9 26.3 ± 4.8 4.5 ± 1.9 Southeast 78.4 ± 5.2 16.5 ± 3.7 4.9 ± 2.1 0.2 ± 0.1 Southwest 44.4 ± 7.1 22.8 ± 1.1 32.5 ± 9.6 0.2 ± 0.1 Statewide 42.4 ± 10.1 16.5 ± 1.1 38.9 ± 9.6 2.2 ± 0.1 A National Land Cover Data Classification System (http://www.mrlc.gov/index.asp). Natural and agricultural cover types exclude developed areas (commercial, residential, and recreational sites) and barren ground (sand, pavement), which represent by region the following mean and (range) of values: Central = 20% (7–60); Northwest = 11% (7–26); Northeast = 13% (8–20); Southeast = 9% (8–14); and Southwest = 14% (6–22). BPredominantly (> 95%) deciduous forest but including evergreen forest, mixed forest, and early seral stage forest. CPasture and hay but also grassland/herbaceous land cover. DCultivated crops and orchards. EBoth woody and herbaceous/emergent wetlands. Northeastern Naturalist 214 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 sampled because: a) they are regularly used in published habitat descriptions and are relevant to habitat preferences of small mammals; b) they represent the most-common, non-cultivated habitat types on our study areas; and c) they are well known and are easily identified for reference in future studies. We did not place transects in cultivated areas because they are disturbed and thus seasonally variable in development during the field seasons (June–December) of this study. Also, coverage of this land-use type within study areas varied widely between regions (Table 1), and allocation of transects to cultivated areas would have reduced sampling effort in natural habitats expected to harbor a higher diversity of small mammals. Grassland was agricultural land used for hay production or livestock grazing. Restored prairie varied from sites seeded predominately with Andropogon gerardi (Big Bluestem) and Sorgastrum nutans (Indian Grass) to areas more fully restored to a mixture of grass and prairie forbs. Wetland transects were placed in wet meadows, typically within 10 m of standing water. Small-mammal collection records for these transects were grouped for analysis in the prairie/wetland category. The number of transects placed within each of the 4 habitat types per study area generally reflected their relative coverage, e.g., the number of woodland transects in the northeast and southeast regions (Tables 1, 2). However, transect placement was not proportional to habitat availability in the Southwest Region, where 30 of 48 transects were placed in grassland and oldfield communities, preferred habitats of Microtus ochrogaster (Prairie Voles) and Reithrodontomys humulis (Eastern Harvest Mice), the two species of interest in that region. Methods Small-mammal sampling We designated ten species of interest in the design of this study, 2 in each of the 5 regions of Ohio (Table 3). These were species that had not been captured recently or were otherwise thought to be rare or of restricted distribution in Ohio based Table 2. Trapping effort in each of 5 regions of Ohio and habitat type is expressed as the number of transects of 50 traps each that were established and the related number of trap nights. Regions of Ohio Central Northwest Northeast Southeast Southwest Total Transects Woodland 16 14 28 25 12 95 Grassland 16 18 14 15 16 79 Oldfield 8 11 3 1 14 37 Prairie and wetlandA 8 13 3 7 6 37 Total transects 48 56B 48 48 48 248 Total trap nightsC 8400D 8400 7200 7200 7200 38,400 AIncludes 6 transects in edge habitat. BAn additional study area was trapped in the northwest region. CTrap nights per transect per study area is the product of 50 traps x 3 nights = 150, yielding 1200 TN per study area. DA study area in Franklin County was trapped for two 3-day sessions, totaling 2400 trap nights. Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 215 Table 3. Non-volant small mammals of Ohio, their historic relative abundance (based on the OWD), and numbers of each species captured in each of the 5 regions of Ohio and state-wide (Total). % = Percent relative abundanceA . Species richness was lowest in the northwest and highest in the southwest region, but regional differences in mean (± SE) abundance (captures/100TN) and the Shannon-Weiner index (H') were not significant (P > 0.05). Species Common nameB % Central NW NE SE SW Total Sorex cinereus (Kerr) Masked Shrew 7.1 87 92 101 39 23 342 Sorex fumeus Miller Smoky Shrew 7.5 0 0 5 6 2 13 Sorex hoyi Baird Pygmy Shrew 0.1 0 0 0 2 0 2 Blarina brevicauda (Say) Short-tailed Shrew 15.5 124 97 90 77 66 454 Cryptotis parva (Say) Least Shrew 1.1 0 0 0 0 6 6 Peromyscus leucopus (Rafinesque) White-footed Mouse 21.3 163 128 127 122 79 619 Peromyscus maniculatus (Wagner) Deer Mouse 13.7 9 7 0 0 3 19 Reithrodontomys humulis (Aububon and Bachmann) Eastern Harvest Mouse 0.3 0 0 0 0 0 0 Myodes gapperi (Vigors) Southern Red-backed Vole 0.05 0 0 0 0 0 0 Microtus pennsylvanicus (Ord) Meadow Vole 17.2 196 178 131 76 88 669 Microtus ochrogaster (Wagner) Prairie Vole 4.8 0 0 0 0 2 2 Microtus pinetorum (LeConte) Woodland Vole 2.9 0 0 1 3 1 5 Synaptomys cooperi Baird Southern Bog Lemming 2.4 1 0 0 2 0 3 Mus musculus Linnaeus House Mouse 3.2 1 0 0 0 1 2 Zapus hudsonius (Zimmermann) Meadow Jumping Mouse 2.6 1 3 1 4 4 13 Napaeozapus insignis (Miller) Woodland Jumping Mouse 0.5 0 0 1 0 0 1 Totals 100 582 505 457 331 275 2150C Captures/100 TND 6.9 6.0 6.3 4.6 3.8 5.6 ± 1.4 ± 0.9 ± 1.4 ± 1.0 ± 0.7 ± 0.6 Species richness 8 6 8 9 11 14 Species diversity (H')D 1.27 1.25 1.16 1.30 1.40 1.27 APercentage of 13,899 specimens in Ohio Wildlife Database represented by each of 16 species of small mammals, i.e., shrews and rodents less than 100 g in body mass. BCommon names of 8 species of interest and the number captured in their designated region are boldfaced. The other 2 species of interest in this study, i.e., Condylura cristata (L.) (Star-nosed Mole) and Spermophilus tridecemlineatus (Mitchill) (Thirteen-lined Ground Squirrel), were not included in this analysis because they were not sampled effectively with the 4 trap types deployed on the standard transect. Capture data for these species are reported separately in the text. CTotal = 2176 with an additional 26 mammals (9 species) that were captured but not included in the analysis: 1 Perascalops breweri (Bachman) (Hairytailed Mole), 2 Scalopus aquaticus (Rafinesque) (Eastern Mole), 3 Condylura cristata (L.), 1 Sylvilagus floridanus (J.A. Allen) (Eastern Cottontail), 5 Tamias striatus (L.) (Eastern Chipmunk), 9 Spermophilus tridecemlineatus, 1 Tamiasciurus hudsonicus (Erxleben) (Red Squirrel), 1 Mustela nivalis L. (Least Weasel), 1 Mustela frenata Lichtenstein (Long-tailed Weasel), and 2 unknown (heavily scavenged and not kept). DCaptures/100 TN (trap nights) and H' (Shannon-Weiner index) were calculated per study area and averaged for presentation by region. ANOVA for captures/ 100 TN and H' among regions was computed with data from 6 study areas per region. Northeastern Naturalist 216 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 on published data (e.g., Gottschang 1981) and/or low representation in the Ohio Wildlife Database (OWD), which contains the electronic records of nearly 22,500 voucher specimens of mammals collected in Ohio during the last century and currently held in 29 collections and museums throughout North America (Harder 1997). Most specimens in the OWD were collected during the 7 decades between 1916 and 1985. The relative abundance of 16 non-volant small-mammal species (i.e., shrews and rodents with a body mass less than 100 g) in the OWD was a reference for planning this study and evaluation of the results. From 17 June–14 December 2004 and 22 June–23 November 2005, we sampled 31 study areas for small mammals: 7 in the northwest region and 6 in each of the other regions (Table 2). We placed the first 2 study areas trapped in each region in locations where the designated species of interest had been most recently captured. Sites for the other 4 study areas per region were selected to provide wide distribution of study areas and broad coverage of each region. We allocated transects on all study areas to sample the 4 designated habitat types rather than to focus on the specialized habitat of any given species. We sampled one new study area per week (Fig.1). To avoid a seasonal bias in allocation of sampling effort to regions of the state, we rotated regions in establishment of successive study areas during the two field seasons. The trapping-transect design used in this study was intended to expose small mammals to a wide variety of traps so as to maximize species richness of the capture. We established each transect with randomized start points in 1 of 4 habitat types, set 10 trap stations at 10-m intervals by pacing on a compass bearing, and recorded the location of each 100-m transect in UTM coordinates at stations 1 and 10 (Fig. 1). We set one Sherman live trap (8 x 8 x 22 cm), a Victor mouse snap trap, and a Museum Special snap trap at each of 10 stations. Traps were set in surface runways when encountered within approximately 2 m of a station. We baited traps with a mixture of rolled oats and peanut butter beginning with the first evening of trapping. We installed five arrays of pitfall traps (4 traps per array) along each transect, one between every other live-snap trap station for a total of 50 traps per transect. Holes for pitfall traps were dug with a golf-cup cutter and lined with a plastic cup (12 cm wide x 15 cm deep). We set the lip of the cup flush with the ground surface and filled the cup with water to a depth of about 8 cm. Four of the 5 arrays of pitfall traps per transect were set with drift fences (0.15 m high x 2.5 m long and anchored with 20-cm gutter spikes), between the center pitfall trap and each of 3 peripheral traps, which were separated by approximately 4.4 m in the y-shaped array (Fig. 1). We followed the recommendation of Kirkland and Sheppard (1994) for use of y-shaped arrays with drift fences in pitfall trapping, but with a design more similar to that used by Kalko and Handley (1993). They used pitfall cups of smaller diameter (11 or 15 cm) with one set in the center and 2 at the ends of 1.2-m drift fences. Although, installation of pitfall traps is time consuming (about 3-person hrs per transect), pitfalls are the most efficient trap type for capturing shrews, and they provide a more accurate estimate of relative abundance of small mammals than Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 217 other trap types (Caceres et al. 2011, Williams and Braun 1983). Traps were set and checked for 3 days, yielding a total of 1200 trap nights per study area (Table 2). We recorded species, age class (adult/juvenile), sex, reproductive status, and capture location for each animal captured. Animals captured in live traps were released unless they could not be readily identified to species, in which case they were euthanized by cervical dislocation. All procedures involving live mammals were approved by the Institutional Animal Care and Use Committee at The Ohio State University. Rarefaction analysis This study was designed to maximize species richness of small mammals captured through use of diverse trap types (4 in all) and intense trapping effort. We assessed the adequacy of this effort to detect uncommon species by constructing sample-based rarefaction curves and their confidence intervals (95% C.I.) (Colwell et al. 2004, 2012) and Chao2 estimates of species richness (95% C.I.) using EstimateS version 8.2 (Colwell 2005) for each study area. Sample-based rarefaction curves are expected species accumulation curves that we calculated using analytical formulae in Colwell et al. (2004). These curves represent the expected number of species captured for a given number of samples, given the set of species captured over all samples in a given study area. Chao2 values are estimates of total species richness, potentially including species that were not captured in any sample. Means and confidence intervals reported were obtained from 50 randomizations of sample order. Here we used the Chao2 estimates to assess the adequacy of sampling effort, rather than as richness estimates per se, by comparing the observed number of species with the estimated richness. Note: the lower 95% C.I. of the Chao2 estimate cannot be less than the observed number of species (Chao 1987). We considered the likelihood of capturing at least 1 additional species, given increased sampling effort, to be low, if the difference between the observed number of species in a study area and the upper 95% C.I. for the Chao2 estimate for that study area was less than 1. Laboratory analyses We retained a representative sample (756 specimens) of the capture for identification of species and preparation of voucher specimens including all 355 long-tailed shrews (Sorex spp.), 92 Peromyscus spp., and 183 Microtus spp. We deposited voucher specimens (totaling 712) in the Museum of Biological Diversity, The Ohio State University, Columbus, OH. Most specimens of Peromyscus spp. (about 90%) examined in the field were identified as White-footed Mice, based on their capture in woodland habitat and having a relatively long non-bicolored tail. The sample of 92 Peromyscus spp. retained for laboratory examination was biased. That is, it included all specimens (n = 25) that showed any indication (i.e., captured in non-forest habitat and with short, bicolored tail) of being Peromyscus maniculatus (Deer Mouse). Also, we retained for further study all vole (Microtus spp.) specimens that showed any indication (pelage and 5 planter tubercles) of being Prairie Vole. The results of Henterly et al. (2011) suggest that very few specimens of Prairie Vole would have been Northeastern Naturalist 218 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 overlooked with this approach. Only 5% (2 of 41) of voles with a Prairie Vole avprla genotype in their study were identified as Meadow Vole when examined once in the field, whereas the converse was true for 29% of voles with a Meadow Vole avpr1a genotype. Mammal carcasses were sprayed with ethanol, placed separately in zipseal plastic bags, chilled on ice in the field, and stored at -10 ºC until they were thawed under ultraviolet light in a fume hood for processing. Small mammals retained as voucher specimens were identified by cranial and dental features of cleaned skulls. Because identification of White-footed Mice and Deer Mice, based on external features or cranial/dental morphology, is subject to substantial error, electrophoresis of salivary amylase is widely recognized as the only reliable method for identification of the two species collected in Ohio (Lindquist et al. 2003, Rich et al. 1996). We compared preliminary identification of White-footed Mice and Deer Mice (based on a sharply bicolored tail and a low taillength- to-total-body-length ratio) to that obtained from acrylamide gel electrophoresis of salivary amylase, using the procedures of Bruseo et al. (1999). We collected saliva by flushing the oral cavity of carcasses with 100 μl of distilled water, which was collected in a microfuge tube and stored at -10 °C until thawed and diluted (1:4) for electrophoretic analysis. Data analysis We expressed relative abundance of small mammals as captures per 100 trap nights (TN) per study area. The Shannon-Weiner index (H') was calculated as an expression of species diversity where H' = -Σ (pi)(ln pi), wherein pi is the proportion of the total sample that belongs to the ith species and ln is the natural log. We calculated captures/100 TN and H' per study area. H' were normally distributed (Shapiro-Wilk test: W = 0.91, df = 31, P > 0.8). Captures/100 TN were square-root transformed to achieve normality (Shapiro-Wilk test on transformed data: W = 0.99, df = 31, P > 0.9). We performed one-way analysis of variance (ANOVA) for differences in captures/100 TN and H' among regions with data from 6 study areas per region, and used Tukey’s post hoc test to check for differences between groups where appropriate. For each of the 4 most abundant species, we tested effects of habitat and region on captures with a generalized linear mixed model (GLMM) using a Poisson distribution and log link. We included captures on each transect as the dependent variable, habitat and region as a fixed effect, and study area nested in region as a random effect. We tested habitat and region effects only for the 4 most abundant species (97% of all captures) because sample sizes for other species (generally less than 10) were too small for meaningful analysis. Post-hoc pairwise comparisons were corrected for multiple testing using a sequential Bonferroni correction. We expressed mean values as a standard error of the mean (SE) and considered differences statistically significant when P < 0.05. Results We captured a total of 2150 small mammals during the two-year study, in which traps were set for 3 nights on each of 31 study areas statewide yielding a total effort Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 219 of 38,400 trap nights. Although 14 of Ohio’s 16 species of small mammals were captured, all but 7 were observed in low numbers (less than 10; Table 3), and 97% of the total capture was comprised of just 4 species: Meadow Vole (31%), White-footed Mouse (29%), Short-tailed Shrew (21%), and Masked Shrew (16%). Estimates of species richness On only 3 study areas (1 in each of central, northwest, and southwest regions) did the mean Chao2 estimate exceed the observed number of species by more than 1, and the difference between the observed number of species and the upper 95% C.I. of the Chao2 richness estimate was less than 1 on 20 of 31 study areas (Fig. 2). This difference threshold was exceeded in 1 study area in the central region, 2 study areas in the northeast and southeast regions, and 3 study areas in the southwest and northwest regions. These results indicate that increased sampling effort in days or transects would not have increased the number of species detected in most study areas, which is in accord with the findings of Conard et al. (2008), who found that 3 nights of trapping are required to reach a stable estimate of species richness at the upper end of the range of trap densities tested (9–144 trap stations/ha). Our sampling with 4 types of trap, Figure 2. Number of species trapped (black dots) and mean Chao2 estimate of species richness (open dots, where different from observed) for each study area, grouped by region of Ohio. Species richness was estimated using the Chao2 estimate (Colwell et al. 2004) and calculated in EstimateS Version 8.2 (Colwell 2005). Error bars for species richness estimates are 95% CI for the mean Chao2 estimates. Northeastern Naturalist 220 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 including 20 pitfall traps per transect, was intended to maximize species richness of the capture. Caceres et al. (2011) concluded that pitfall traps were particularly effective in this regard, capturing all 14 species known to inhabit their study area, while Sherman and wire traps captured only 8 (57%). Abundance and diversity: Variation by region and habitat Region. Overall abundance of small mammals (all species) varied widely among study areas from 13.2 captures per 100 TN on a study area in the northeast region to 2.1/100 TN on one in the southwest region. However, regional differences in total small-mammal abundance (Table 3) were not significant overall (ANOVA: F = 1.24; df = 4, 26; P > 0.3). Also, regional differences among the 4 most abundant species (Table 4) were not significant (GLMM: Meadow Vole: F = 0.971; df = 4, 234; P = 0.42; White-footed Mouse: F = 0.809; df = 4, 234; P = 0.52; Short-tailed Shrew: F = 0.665; df = 4, 234; P = 0.62; and Masked Shrew: F = 1.966; df = 4, 234; P = 0.10). Although not significant (P = 0.10), regional variation in abundance of Masked Shrews was high, captures/100 TN being 5 times higher in the Northeast than in the southwest region (Table 4). Although small-mammal abundance in the two northern regions was about 45% higher (P > 0.05) than in the 2 southern regions, the highest species richness for the designated small mammals in this study (11) was recorded in the southwest region (Table 3). The southwest region also had the distinction of being the only region home to all 7 of the most abundant species (Table 4) and to one of the rarest species, the Prairie Vole (Table 3). However, regional differences in species diversity (H') were not significant (ANOVA: F = 1.44; df = 4, 26; P > 0.2; Table 3). It should be noted that our regional comparisons might be biased or compromised because placement of the first 2 study areas per region were associated with the most recent capture locations for different species of interest across regions. Habitat. The relative abundance of small mammals (all species, captures/100 TN) in each of the 4 habitat types viewed statewide (woodland = 5.8, grassland = 6.5, oldfield = 5.4, prairie/wetland 6.0) was similar (P > 0.05). This overall similarity in small-mammal abundance by habitat reflects the counterbalancing effect of habitat associations of the most abundant species (Table 5). The effect of habitat was significant for Masked Shrews, (GLMM—pitfall only: F = 11.380; df = 3, 234; P < 0.001; all traps: F = 11.288; df = 3, 234; P < 0.001), White-footed Mice (F = 100.053; df = 3, 234; P < 0.001), and Meadow Voles (F = 45.210; df = 3, 234; P less than 0.001), but not for Short-tailed Shrews (F = 2.504; df = 3, 234; P = 0.06). Masked Shrews were most abundant in restored prairie/wetland (P < 0.05), and White- Footed Mice were captured most frequently in woodlands (P < 0.5). By contrast, Meadow Voles were most abundant in grassland-pasture habitat and infrequently captured in woodlands (P < 0.05; Table 5). Species of interest All 10 species of interest were captured infrequently (less than 10 captures), and 3 of them (Prairie Vole, Microtus pinetorum [Woodland Vole], and Synaptomys cooperi Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 221 Table 4. Mean (± SE)A relative abundance (captures/100 TN) of the 7 most abundant small mammals in 5 regions of Ohio and the number of study areas in which each species was captured. Regional dif ferences in abundance were not significant (P > 0.05). Species Central Northwest Northeast Southeast Southwest All regions Study areas Masked ShrewB All traps: 1.06 ± 0.46 1.10 ± 0.28 1.40 ± 0.42 0.51 ± 0.21 0.28 ± 0.11 0.88 ± 0.16 28 Pitfall traps only: 2.43 ± 1.07 2.29 ± 0.62 3.37 ± 0.98 1.17 ± 0.48 0.67 ± 0.28 2.01 ± 0.36 Smoky Shrew 0.00 0.00 0.07 ± 0.05 0.08 ± 0.06 0.03 ± 0.03 0.04 ± 0.02 5 Short-tailed Shrew 1.48 ± 0 .38 1.14 ± 0.24 1.25 ± 0.27 1.09 ± 0.32 0.86 ± 0.20 1.17 ± 0.13 31 White-footed Mouse 1.80 ± 0.44 1.50 ± 0.39 1.76 ± 0.38 1.69 ± 0.34 1.11 ± 0.32 1.57 ± 0.17 31 Deer Mouse 0.12 ± 0.03 0.08 ± 0.04 0.00 0.00 0.04 ± 0.40 0.05 ± 0.02 9 Meadow Vole 2.38 ± 0.51 2.12 ± 0.70 1.82 ± 1.41 1.02 ± 0.67 1.23 ± 0.33 1.75 ± 0.34 28 Meadow Jumping Mouse 0.01 ± 0.01 0.04 ± 0.02 0.01 ± 0.01 0.06 ± 0.03 0.06 ± 0.04 0.04 ± 0.01 10 ACaptures/100 TN for each species were calculated for each of 31 study areas and averaged by region. BMost (90%) were captured in pitfall traps; so, abundance was calculated for all traps (50/transect) and for pitfall traps alone (20/transect). Northeastern Naturalist 222 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 [Southern Bog Lemming]) plus Sorex fumeus (Smoky Shrew) and Deer Mouse were captured in remarkably low numbers (Table 3). Two others, the Eastern Harvest Mouse and Southern Red-backed Vole, were not captured at all during the course of our 2-year study. Only 19 of the 92 specimens (20.6%) of Peromyscus spp. retained for laboratory examination of skulls and salivary amylase were ultimately identified as Deer Mice. The sample of Peromyscus spp. retained for laboratory examination included all specimens (n = 25) that showed any indication of being Deer Mice (see Methods). These results indicate that only about 3% (19 of all 638 Peromyscus spp. captured in this study; Table 3) were Deer Mice. Habitat associations support this conclusion; White-footed Mice were most abundant in forest habitat (Table 5), and most of the Deer Mice (14/19) were collected in grassland or ol dfield habitats. Condylura cristata (Star-nosed Mole) and Spermophilus tridecemlineatus (Thirteen-lined Ground Squirrel) were designated species of interest for this study because of their limited distribution and low relative abundance in the OWD. However, they were not susceptible to capture in the array of traps deployed on our sampling transects and, therefore, they were not included in Table 3 and related analyses. We captured two Star-nosed Moles in Coshocton County (northern Southeastern Region) and a third in Williams County (northwest corner of the state) All 3 were captured in pitfall traps set in wetland habitat. Thirteen-lined Ground Squirrels reach the eastern limit of their continental distribution in central Ohio, where the existence and size of colonies vary widely over time. With the aid of local residents, we located robust colonies and established study areas in Clark and Ross counties (southwest region) and in Licking and Pickaway counties (central region). Colonies were not identified in Muskingum County, which is the eastern-most county in Ohio from which specimens have been collected in the past. We set rat traps (20 per study area) at approximately 20-m intervals near apparent burrow entrances for 3 nights. We caught three squirrels in Licking County and 2 at each of the other 3 county sites. Table 5. Average (± SE) captures/100 TN of the 4 most abundant small mammals in each of the 4 major habitat types sampled in the study. Within a species, values followed by different letters are significantly different (P < 0.05) from one anotherA. Habitat types Species Woodland Oldfield Grassland Prairie/ wetland Masked ShrewB Pitfall traps only 1.96 ± 0.38a 1.80 ± 0.49a 1.60 ± 0.31a 4.61 ± 1.04b All traps: 0.79 ± 0.15a 0.94 ± 0.26a 0.74 ± 0.13a 1.98 ± 0.45b Short-tailed Shrew 1.39 ± 0.20a 1.30 ± 0.39a 1.25 ± 0.17a 0.76 ± 0.19a White-footed Mouse 3.28 ± 0.28a 0.94 ± 0.31b 0.32 ± 0.13c 1.24 ± 0.41d Meadow Vole 0.07 ± 0.06a 1.96 ± 0.42b 3.94 ± 0.57c 2.00 ± 0.39b ATotal captures per transect (150 TN) compared among habitats using a generalized linear mixed model with a Poisson distribution and log-link, with study area as a random effect. BAbundance for this species was calculated for all traps (50/transect) and for pitfall traps alone (20/ transect). Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 223 Discussion To our knowledge, this is the first large-scale, systematic study of its kind in North America in several decades. Early information on mammals in Ohio dates from the publications of Brayton (1882), Kirtland (1838) and most notably Enders (1930), who trapped mammals on 28 study areas in 10 counties in eastern and southern Ohio, and the comprehensive review of museum records by Bole and Moulthrop (1942). The North American Census of Small Mammals, conducted during 1948–1959, includes reports from cooperators in Ohio (Calhoun 1949). Jorgensen (2004) examined 70 studies of mammalian diversity and microhabitat published between 1969 and 2000 and found that most projects (>50%) involved only 1 or 2 vegetation types, 8 or fewer species, and relatively small spatial scales. Greatest lineal distances between study sites in North American studies are typically less than 10 km, and only a few involve distances in the range of 50 to 150 km (Cudmore and Whitaker 1984, Iverson et al. 1967, Rickard 1960) or reach 340 km as in Garneau et al. (2012). The 31 study areas of this investigation were distributed statewide, representing north–south and east–west lineal distances of 335 and 350 km, respectively (Fig. 1). This wide geographic distribution of study sites in 4 major habitat types increases the relevance of our results to concerns regarding the abundance and diversity of small mammals in other states in the upper Midwest. Habitat and small-mammal abundance Three of the 4 most common species showed clear habitat associations (Table 5). The Short-tailed Shrew was the exception with similar frequency of capture in each of the habitat types. Although DeCapita and Bookhout (1975) found this shrew more often than expected in oldfield-pine habitat in southeastern Ohio, our finding is in accord with conclusions of Bowman et al. (2000), Iverson et al. (1967), and Osbourne et al. (2005) and supports consensus that the Short-tailed Shrew is a habitat generalist. White-footed Mouse. White-footed Mice were found on all 31 study areas, and although they were well represented in all habitats (except grasslands), their abundance was highest (P < 0.05) on wooded sites, 2 to 10 times higher than in the other 3 habitats (Table 5). White-footed Mice occupy diverse habitats throughout their range in North America but favor wooded areas, even in tallgrass prairie (Matlack et al. 2008), with vertical density and structural complexity (Drickamer 1990, M’Closkey 1975). Studies of White-footed Mice in Ohio have found highest densities in the structurally complex edge habitat of smaller forest patches (Anderson and Meikle 2006) and have demonstrated the importance the annual mast crop in determining the size of the reproducing population in early spring (Vessey and Vessey 2007). Meadow Vole. Statewide, Meadow Voles were captured more frequently than any other small mammal, except Masked Shrews when calculated for pitfall traps alone. Being grazers, Meadow Voles are largely restricted to habitats with grass or herbaceous ground cover (Getz 1985), which is reflected in their highest abundance (P < 0.05) on transects in grassland habitats and virtual absence from woodland Northeastern Naturalist 224 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 sites (Table 5). Meadow Vole abundance was highly variable, particularly in the northeast region (Table 4), where the number of voles captured per study area varied from 2 to 105. Spatial variation in vole abundance may be related to sitespecific conditions, such as cover and food (Getz 1985, Getz et al. 2001), whereas interaction between food supply and predation appears to be a key factor affecting population cycles in voles (Krebs 1996). Based on a 25-year demographic study of Prairie Voles (Getz et al. 2006a) and Meadow Voles (Getz et al. 2006b), the authors concluded that population fluctuations were initiated by the net effects of periodic relaxation of predation pressure. Masked Shrew. Because nearly all Masked Shrews (90%) were captured in pitfall traps (the highest trap specificity of any species), the abundance of this shrew should logically be calculated for pitfall traps alone (20/transect), and by this measure, the abundance of Masked Shrews (2.01/100 TN, in all regions) ranked higher than that for all other species (Table 4). However, it should be noted that because most Masked Shrews were captured in pitfall traps, which can function as repeating traps (i.e., capturing more than one animal per trap night), data on captures/100 TN might under-represent the trapping effort (TN) and thus overexpress the relative abundance of this species. Masked Shrews were trapped in 16 counties where they had not been previously recorded, and 5 of those counties were well outside the previously published range in northern and south-central Ohio (Gottschang 1981). Masked Shrews were also the most abundant small mammal in several other study sites in North America including Manitoba (Wrigley et al. 1979), Prince Edward Island (Hartling and Silva 2003), southern Pennsylvania (Kirkland and Findley 1999), and West Virginia (Osbourne et al. 2005). Masked Shrews are found in a variety of habitats that provide a mesic microclimate (Brooks and Doyle 2001, Menzel et al. 1999, Wrigley et al. 1979). Our data suggest that restored prairie and wetlands provide favorable, mesic habitat for the Masked Shrew in Ohio where it reaches the southern limit of its distribution in eastern North America, outside Appalachia. Masked Shrews have high metabolic and water-turnover rates (Poppitt et al. 1993, Speakman 1997), which appear to limit their southern distribution into warmer, drier climates of North America (Churchfield 1990, Ochocinska and Taylor 2005) and favors their southernmost distribution in cool, moist locations at high elevations of the Appalachian Mountains (Brannon 2002, Ford et al. 2001). Similarly in this study, abundance of Masked Shrews was 5 times higher in the Northeast than in the Southwest Region (P = 0 .10; Table 4). Northeast Ohio is characterized by lower mean summer temperatures than southwest Ohio (about 3.6 °C less) (NOAA 2010), which might contribute to cool mesic habitats favorable to shrew populations during the growing season. Smoky Shrew. We captured 13 Smoky Shrews, all on study areas in the Glaciated Plateau or Unglaciated Appalachian Plateau (Fig. 1), confirming the observations of Gottschang (1981). All Smoky Shrews were captured in pitfall traps set in wooded areas, and although our sample is too small for statistical analysis, this pattern is consistent with a specialized habitat preference for shady, damp woods with leaf litter and decomposing logs found in Ohio (Gottschang 1981), North Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 225 Carolina (Brannon 2002), and Indiana (Cudmore and Whitaker 1984). The relative abundance (captures/100 pitfall TN in woodland habitat) of Smoky Shrews in our study (0.35) is similar to that (0.36) recorded in southern Indiana by Cudmore and Whitaker (1984), who set pitfalls (1000-ml plastic beakers) along natural obstructions (logs, rock faces) in suitable forest habitat, but lower than the 5.0 recorded by Brannon (2002) in North Carolina and the 1.50 recorded by Kirkland and Findley (1999) in southern Pennsylvania. Both of these latter studies employed y-shaped pitfall arrays in lowland forest habitat. Additional, targeted surveys designed to sample the restricted habitat of Smoky Shrews with pitfall trap arrays are needed to fully assess the status of the species in Ohio. Small-mammal diversity Statewide species diversity (H' = 1.27) was lower than the H' of 1.78 for small mammals trapped in riparian forests in West Virginia (Osbourne et al. 2005), and the H' of 1.57 from a sample of 265 trapped in forests southern Pennsylvania (Kirkland and Findley 1999), but similar to the H' of 1.33 from small mammals trapped in the Catskill region of New York (Garneau et al. 2012). Status of small mammals in Ohio This study was designed to assess the status of 16 species of small mammals in Ohio including the abundance and distribution of 10 species of interest, which may now be placed in 1 of 2 groups. The first group of 5 (Least Shrew, Thirteen-lined Ground Squirrel, Sorex hoyi [Pygmy Shrew], Star-nosed Mole, and Napaeozapus insignis [Woodland Jumping Mouse]) were, prior to this study, considered rare, uncommon, or restricted in distribution (according to Gottschang 1981, Svendson 1976, and their relative abundance in the OWD), and the last 3 were listed as species of concern by the ODW, Department of Natural Resources (ODW 2012). No evidence was obtained in this study that would suggest a change in the conservation status for any of the 5 species in this group. The second group of 5 species of interest includes 2 (Eastern Harvest Mouse and Southern Red-backed Vole) that were, prior to this study, thought to be very rare, if not extirpated from Ohio, whereas the other 3 (Prairie Vole, Woodland Vole, and Southern Bog Lemming) were considered to be of moderate abundance (Gottschang 1981), each comprising >2% of all small-mammal specimens in the OWD. Southern Red-backed Vole. The Southern Red-backed Vole was last recorded in far eastern Ohio (Jefferson County) in 1960 (OWD), and despite our trapping in the general area of the last 2 known capture locations, none was captured in this study. In fact, only 7 specimens have ever been collected in Ohio, and 6 of these were captured in 1926–1929 from sites in eastern Ashtabula County that were subsequently flooded by the Pymatuning Reservoir (Bole and Moulthrop 1942). Although the Southern Red-backed Vole is abundant in northeastern states bordering Ohio, it is listed as extirpated from Ohio (ODW 2012), which is in accord with the conclusion of Bole and Moulthrop (1942) and Gottschang (1981). Eastern Harvest Mouse. The Eastern Harvest Mouse was not captured in this study, but it does not appear to be extirpated from Ohio. It was previously reported Northeastern Naturalist 226 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 in 9 counties in southern Ohio, and it was captured on 2 of 5 units during a year-long (12-month) survey of the Hopewell Culture Historic National Park near Chillicothe in south-central Ohio (Vick 2004). Our procedures probably were not sufficiently targeted or intense to detect this species on most study areas. However, this species is rare, spotty in distribution, and not easily trapped (Gottschang 1965), and Ohio lies at the northern limit of its distribution. It was first captured in Adams County (borders the Ohio River) in 1929 (Bole 1932), and prior to the recent captures near Chillicothe, captures had not been reported since 1971. It currently is listed as threatened in Ohio (ODW 2012). Woodland Vole. Even though 65 transects (9750 TN) were placed in woodland habitat in the northeast, southeast, and southwest regions, only 5 Woodland Voles were captured (Table 3). We might have captured more had our procedures been targeted for this species, e.g., setting pitfall traps in subsurface burrows or along natural obstructions, as in the study of Cudmore and Whitaker (1984), who captured 69 Woodland Voles (in 29,967 TN) in a study of shrews in southern Indiana. Woodland Voles were not captured in the central or northwest regions, which supports Gottschang’s (1981) conclusion that this vole is absent from northwestern Ohio. Although Woodland Voles are widely distributed throughout eastern United States, the distributional pattern in Ohio might be related to low forest cover and extensive cultivation in the northwest region (Table 1) or perhaps soil texture as in orchards in Pennsylvania (Fisher and Anthony 1980). Southern Bog Lemming. The status of this species in Ohio remains unclear. The OWD contains a moderate number of records of Southern Bog Lemmings (328 or 2.4% of total small mammals), with captures recorded from nearly half (40 of 88) of the counties and all 5 regions of the state, but despite our trapping on 79 transects (11,850 TN) in grassland habitat (favored by bog lemmings in the eastern Midwest), we captured only 3 individuals. The Southern Bog Lemming is known, in the eastern part of its range, to be rare or at least spotty in distribution and difficult to capture (Gottschang 1981, Linzey 1983). We set our traps in runways, within constraints of our transect protocol, but we seldom observed the bright green droppings characteristic of this species. No doubt more targeted procedures would have captured more Southern Bog Lemmings, but the low number recorded in our study might also reflect competitive exclusion by Meadow Voles. Linzey (1984) provided convincing evidence of this in Virginia, and a decline of Southern Bog Lemmings was associated with a range extension of Meadow Voles in Kentucky (Krupa and Haskins 1996). Thus, future studies should not only address the status of this species of concern in Ohio but also explore its ecological relationship with Meadow Voles. Prairie Vole. Perhaps the most striking manifestation of low species diversity of small mammals observed in this study is the dominance of Meadow Voles and White-footed Mice, which represent 96.6% of all rodent captures. The Prairie Vole is the seventh-most abundant species of small mammal in the OWD with records from 30 counties, and it had once been considered to be fairly common in the southern half of Ohio. In fact, it was equal in abundance (377 captures) to Meadow Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 227 Voles (303 captures) in one study area near Dayton (DeCoursey 1957). However, despite our deployment of a large number of transects in grassland and prairie/wetland habitats in the central and southwest regions, totaling 6900 TN, only 2 Prairie Voles were captured during the entire study, both in Adams County in the southwest region. By comparison, 35 Prairie Voles (and 7 Southern Bog Lemmings) were captured in lines of snap-back mouse traps (5538 TN) set in dry oldfield habitat in western Indiana (Whitaker et al. 2007). Prairie Voles reach the eastern limit of their distribution in eastern Ohio, western Pennsylvania, and West Virginia. Having been first recorded in Ohio in 1921 (Henninger 1921), this vole now appears to persist in very low numbers in Ohio, and it is listed as a species of concern (ODW 2012). Deer Mouse. Our conclusion that only about 3% of all Peromyscus spp. captured in this study were Deer Mice is noteworthy with reference to records in the OWD, wherein 39% (1903 of 4851) of the specimens of Peromyscus spp. were recorded as Deer Mice. We did not sample cultivated areas, and they are known to harbor substantial populations of Deer Mice and Mus musculus (House Mouse) in Illinois (Getz and Brighty 1986) and Indiana (Whitaker 1967), which best explains the capture of only 2 specimens of the latter (Table 3). Whitaker (1967) found the abundance of Deer Mice in several cropland types (wheat, corn, soybean) to be equivalent to or 2–3 times higher than in weedy or grassy fields. Previous studies in Ohio have shown Deer Mice to be relatively abundant in grassland and oldfield habitats (DeCoursey 1957, Gottschang 1981). Birch (1977) found that most (57 of 74) of the Peromyscus spp. captured in oldfield and grassland habitats in south-central Ohio were Deer Mice (rigorously indentified by discriminate analysis of morphometrics). But, despite intensive sampling in grassland and oldfield habitats (11,850 and 5500 TN, respectively), we captured only 19 Deer Mice, and this species was absent in samples collected in 2 of 5 regions (Table 4). Nevertheless, because we did not sample cultivated fields, we cannot estimate the overall relative abundance of Deer Mice in the state. Additional, targeted surveys employing electrophoretic or molecular techniques for identification of Deer Mice are needed to resolve the status of this species in Ohio. Star-nosed Moles and Thirteen-lined Ground Squirrels. Since all 3 Star-nosed Moles captured were caught in wetlands, future efforts to monitor this species of concern (ODW 2012, 2013) should focus on wetland habitat in northern Ohio. Our capture results for Thirteen-lined Ground Squirrels and earlier observations cited in Gottschang (1981) indicate that the eastern range limit of this species in central and southeastern Ohio has not changed appreciably in the last 30–50 years. Summary and Conclusions Twenty-three species of mammals were captured in this study, 14 of which were small mammals (shrews and small rodents less than 100 g in body mass) and the focus of our analysis; 97% of 2150 captures of small mammals was represented by just 4 species: Meadow Vole (31%), White-footed Mouse (29%), Short-tailed Shrew (21%), and Masked Shrew (16%). Species richness varied from 6 in the northwest Northeastern Naturalist 228 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 region to 11 in the southwest region. Rarefaction analysis indicated that additional trapping effort would not have increased the number of species detected in most study areas. Regional differences in abundance of small mammals (captures/100 trap nights) and species diversity (H') were not significant (P > 0.05). However, regionally restricted distributions were confirmed for the Smoky Shrew and Woodland Vole. This study also confirmed well-known habitat preferences for Whitefooted Mice and Meadow Voles and provided the first evidence of the preferred habitat of Masked Shrews in Ohio: restored prairie and wetland. Our systematic transect-sampling procedure might be considered a coarse filter that is well suited to sampling common to moderately abundant mammals (e.g., the 7 species in Table 5, except Smoky Shrew) in major habitat types, but it might not detect or, in particular, effectively estimate relative abundance of rare species, especially habitat specialists or species that do not readily enter traps. This category would include most of our designated species of interest except Deer Mice (absence of samples from cultivated habitat, notwithstanding) and Prairie Voles. Four rodent species of questionable status in Ohio (Eastern Harvest Mouse, Prairie Vole, Southern Bog Lemming, and Deer Mouse) live in grassland or oldfield communities, which suggests the quantity or quality of these habitats in Ohio may be insufficient to support large populations of these species. A meta-analysis of studies of South African grasslands indicated that small-mammal species richness and diversity declines with habitat degradation and that generalist rodent species dominate community numbers at low ecological integrity (Avenant 2011). Changing agricultural land-use patterns in Ohio, such as the 25% reduction of acreage devoted to hay fields and an 11% increase in that used for corn and soybean production during 1973–2004 (NASS 2011), might be associated with a general decline in small-mammal habitat. Additional research is needed to confirm the status of grassland and oldfield species and to reveal the underlying causes of any changes in Ohio as well as in other states in the southern Great Lakes region. Acknowledgments We are grateful to David Lowell and Matthew Carney for their invaluable assistance during 2 rigorous field seasons and Travis Brown for his outstanding contribution to the project in preparation of voucher specimens. We thank Karen Treadway, Amanda Prouty, Brad Ryan, Brad Condo, and Kyle Tolle for technical assistance; Joni Lung, Adam Andrews, Travis Fuller, and Brad Graley for field work; and Jimmy Chiucchi for statistical analysis. Donovan Powers, ODNR, created Figure 1 and provided output from the NLDCL for percentage cover types for the 31 study areas. Christine Anderson reviewed a draft of the manuscript. We relied heavily on Ohio Division of Wildlife personnel and private landowners, too numerous to mention individually, who hosted our fieldwork and assisted in identifying suitable trapping sites as did administrators and field personnel with the Ohio State Parks, Ohio State Forests, Wayne National Forest, Ottawa National Wildlife Refuge, Columbus Franklin County Metro Parks, The Nature Conservancy, and the National Aeronautics and Space Administration, Plum Brook Station. Carolyn Caldwell, ODW, provided administrative support and valuable consultation and encouragement throughout the course of the study. This project was funded by the Ohio Division of Wildlife, ODNR. Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 229 Literature Cited Anderson, C.S., and D.B. Meikle. 2006. Annual changes in structural complexity of understory vegetation and relative abundance of Peromyscus leucopus in fragmented habitats. Acta Theriologica 51:43–51. Avenant, N. 2011. The potential utility of rodents and other small mammals as indicators of ecosystem “integrity” of South African grasslands. Wildlife Research 38:626–639. Bateman, B.L., A.S. Kutt, E P. Vanderduys, and J.E. Kemp. 2010. Small-mammal species richness and abundance along a tropical altitudinal gradient: An Australian example. Journal of Tropical Ecology 26:139–149. Birch, W.L. 1977. Ecological separation of Peromyscus maniculatus bairdii and Peromyscus leucopus noveboracensis (Rodentia) in southcentral Ohio. M.Sc. Thesis. The Ohio State University, Columbus, OH. 86 pp. Bole, B.P., Jr. 1932. An addition to the known list of Ohio mammals. Ohio Journal of Science 32:402. Bole, B.P., Jr., and P. N. Moulthrop. 1942. The recent mammal collection in the Cleveland Museum of Natural History. Scientific Publications of the Cleveland Museum of Natural History 5:83–181. Bowman, J.C., D. Sleep, G.J. Forbes, and M. Edwards. 2000. The association of small mammals with coarse woody debris at log and stand scales. Forest Ecology and Management 129:119–124. Brannon, M.P. 2002. Distribution of Sorex cinereus and S. fumeus on north- and south-facing slopes in the southern Appalachian Mountains. Southeastern Naturalist 1:299–306. Brayton, A.W. 1882. Report on the mammals of Ohio. Report of the Geological Survey of Ohio 4:1–185. Brooks, R.T., and K.L. Doyle. 2001. Shrew species richness and abundance in relation to vernal pond habitat in southern New England. Northeastern Naturalist 8:137–148. Bruseo, J.A., S.H. Vessey, and J.S. Graham. 1999. Discrimination between Peromyscus leucopus noveboracensis and Peromyscus maniculatus nubiterrae in the field. Acta Theriologica 44:151–160. Calhoun, J.B. 1949. North American census of small mammals. Rodent Ecology Project, Johns Hopkins University, Release 2:49–50. Caceres, N.C., R.P. Napoli, and W. Hannibal. 2011. Differential trapping success for small mammals using pitfall and standard cage traps in a woodland savannah region of southwestern Brazil. Mammalia 75:45–52. Caro, T.M., M.J. Kelly, N. Bol, and S. Matola. 2001. Inventorying mammals at multiple sites in the Maya Mountains of Belize. Journal of Mammalogy 82:43–50. Chao, A. 1987. Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43:783–791. Churchfield, S. 1990. The Natural History of Shrews. Christopher Helm, London, UK. 178 pp. Colwell, R.K. 2005. EstimateS: Statistical estimation of species richness and shared species from samples. Version 8.2. Available online at http://purl.oclc.org/estimates. Accessed 22 August 2012. Colwell, R.K., C.X. Mao, and J. Chang. 2004. Interpolating, extrapolating, and comparing incidence-based species-accumulation curves. Ecology 85:2717–2727. Colwell, R.K., A. Chao, N.J. Gotelli, S.Y. Lin, C.X. Mao, R.L. Chazdon, and J.T. Longino. 2012. Models and estimators linking individual-based and sample-based rarefaction, extrapolation, and comparison of assemblages. Journal of Plant Ecology 5:3–21. Northeastern Naturalist 230 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 Conard, J.M., J.A. Baumgardt, P.S. Gipson, and D.P. Althoff. 2008. The influence of trap density and sampling duration on the detection of small-mammal species richness. Acta Theriologica 53:143–156. Cudmore, W.W., and J.O. Whitaker, Jr. 1984. The distribution of the Smoky Shrew, Sorex fumeus, and the Pygmy Shrew, Microsorex hoyi, in Indiana, with notes on the distribution of other shrews. Proceedings of the Indiana Academy of Science 93:469–474. DeCapita, M.E., and T.A. Bookhout. 1975. Small-mammal populations, vegetational cover, and hunting use of an Ohio strip-mined area. Ohio Journal of Science 75:305–313. DeCoursey, G.E., Jr. 1957. Identification, ecology and reproduction of Microtus in Ohio. Journal of Mammalogy 38:44–52. Drickamer, L.C. 1990. Microhabitat preferences of two species of deer mice, Peromyscus, in a northeastern United States deciduous hardwood forest. Acta Theriologica 35:241–252. Enders, R.K. 1930. Some factors influencing the distribution of mammals in Ohio. Occasional Papers, University of Michigan Museum of Zoology 212:1–27. Entwistle, A.C., and P.J. Stephenson. 2002. Small mammals and the conservation agenda, Pp. 119–139, In A.C. Entwistle and N. Dunstone (Eds.). Priorities for the Conservation of Mammalian Diversity. Cambridge University Press, Cambridge, UK. 436 pp. Fisher, A.R., and R.G. Anthony. 1980. The effect of soil texture on distribution of Pine Voles in Pennsylvania orchards. American Midland Naturalist 104:39–46. Ford, W.M., M.A. Menzel, T S. McCay, and J. Laerm. 2001. Contiguous allopatry of the Masked Shrew and Southeastern Shrew in the southern Appalachians: Segregation along an elevational and habitat gradient. Journal of the Elisha Mitchell Science Society 117:20–28. Garneau, D.E., M.E. Lawler, A.S. Rumpf, E.S. Weyburne, T.M. Cuppernull, and A.G. Boe. 2012. Potential effects of beech bark disease on small mammals and invertebrates in northeastern US forests. Northeastern Naturalist 19:391–410. Getz, L.L. 1985. Habitats. Pp. 286–309, In R.H. Tamarin (Ed.). Biology of New World Microtus. Special Publication No. 8. The American Society of Mammalogists. 893 pp. Getz, L.L., and E. Brighty. 1986. Potential effects of small mammals in high-intensity agricultural systems in east-central Illinois, USA. Agriculture, Ecosystems, and Environment 15:39–50. Getz, L.L., J. E. Hoffmann, B. McGuire, and T.W. Dolan. 2001. Twenty-five years of population fluctuations of Microtus ochrogaster and M. pennsylvanicus in 3 habitats in eastcentral Illinois. Journal of Mammalogy 82:22–34. Getz, L.L., M.K. Oli, J.E. Hofmann, and B. McGuire. 2006a. Vole population fluctuations: Factors that initiate and determine intervals between them in Microtus ochrogaster. Journal of Mammalogy 87:387–393. Getz, L.L., M.K. Oli, J.E. Hofmann, and B. McGuire. 2006b. Vole population fluctuations: Factors that initiate and determine intervals between them in Microtus pennsylvanicus. Journal of Mammalogy 87:841–847. Gottschang, J.L. 1965. Winter populations of small mammals in old fields in southwestern Ohio. Journal of Mammalogy 46:44–52. Gottschang, J.L. 1981. A Guide to the Mammals of Ohio. The Ohio State University Press, Columbus, OH. 176 pp. Harder, J.D. 1997. The Ohio Mammals Database: Distribution and history of mammals collected in Ohio. Final Report. Ohio Division of Wildlife, Ohio Department Natural Resources, Columbus, OH. 13 pp. Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 231 Hartling, L., and M. Silva. 2003. Abundance and species richness of shrews within forested habitats on Prince Edward Island. American Midland Naturalist 151:399–407. Henninger, W. F. 1921. Two mammals new for Ohio. Journal of Mammalogy 2:239. Henterly, A.C., K.E. Mabry, N.G. Solomon, and A.S. Chesh. 2011. Comparison of morphological versus molecular characters for discriminating between sympatric Meadow and Prairie Voles. American Midland Naturalist 165:412–420. Iverson, S.L., R.W. Seabloom, and J.M. Hnatiuk. 1967. Small-mammal distributions across the prairie-forest transition of Minnesota and North Dakota. American Midland Naturalist 78:188–197. Jorgensen, E.E. 2004. Small-mammal use of microhabitat reviewed. Journal of Mammalogy 85:531–539. Kalko, E.K.V., and C.O. Handley, Jr. 1993. Comparative studies of small-mammal populations with transects of snap traps and pitfall arrays in southwest Virginia. Virginia Journal of Science 44:3–19. Kirkland, G.L., Jr., and J.S. Findley. 1999. A transcontinental comparison of forest smallmammal assemblages: Northern New Mexico and southern Pennsylvania compared. OIKOS 85: 335–342. Kirkland, G.L., Jr., and P.K. Sheppard. 1994. Proposed standard protocol for sampling small-mammal communities. Pp 277–284, In J.F. Merritt, G.L. Kirkland, Jr., and R.K. Rose (Eds.). Advances in the Biology of Shrews. Carnegie Museum of Natural History, Pittsburgh, PA. 458 pp. Kirtland, J.P. 1838. Report on mammals. Geological Survey of Ohio 2:160–161, 175–177. Krebs, C.J. 1996. Population cycles revisited. Journal of Mammalogy 77:8–24. Krupa, J.J., and K.E. Haskins. 1996. Invasion of the Meadow Vole (Microtus pennsylvanicus) in southeastern Kentucky and its possible impact on the Southern Bog Lemming (Synaptomys cooperi). American Midland Naturalist 135:14–22. Linzey, A.V. 1983. Synaptomys cooperi. Mammalian Species 210:1–5. Linzey, A.V. 1984. Patterns of coexistence in Synaptomys cooperi and Microtus pennsylvanicus. Ecology 65:382–393. Lindquist, E.S., C.F. Aquadro, D. McClearn, and K.J. McGowan. 2003. Field identification of the mice Peromyscus leucopus noveboracensis and P. maniculatus gracilis in central New York. Canadian Field Naturalist 117:184–189. Manley, P.N., M.D. Schlesinger, J.K. Roth, and B. Van Horne. 2005. A field-based evaluation of a presence-absence protocol for monitoring ecoregion-scale biodiversity. Journal of Wildlife Management 69:950–966. Matlack, R.S., D.W. Kaufman, and G.A. Kaufman. 2008. Influence of woody vegetation on small mammals in tallgrass prairie. American Midland Naturalist 160:7–19. M’Closkey, R.T. 1975. Habitat dimensions of white-footed mice, Peromyscus leucopus. American Midland Naturalist 93:158–167. McShea, W. J., J. Pagels, J. Orrock, E. Harper and K. Koy. 2003. Mesic deciduous forest as patches of small-mammal richness within an Appalachian mountain forest. Journal of Mammalogy 84:627–643. Menzel, M.A., W.M. Ford, J. Laerm, and D. Krishon. 1999. Forest to wildlife opening: Habitat-gradient analysis among small mammals in the southern Appalachians. Forest Ecology and Management 114:227–232. National Agricultural Statistics Service (NASS). 2011. Ohio Historical Agricultural Statistics. Available online at http://www.nass.usda.gov/. Accessed 16 June 2011. National Oceanic and Atmospheric Administration (NOAA). 2010. National Climatic Data Center. Available online at http://www.ncdc.noaa.gov/oa/ncdc.html. Accessed 14 December 2010. Northeastern Naturalist 232 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 Vol. 21, No. 2 Ochocinska, D., and J.R.E. Taylor. 2005. Living at the physiological limits: Field and maximum metabolic rates of the Common Shrew (Sorex araneus). Physiological and Biochemical Zoology 78:808–818. Ohio Division of Wildlife (ODW). 2012. Wildlife that are considered to be endangered, threatened, species of concern, special interest, extirpated, or extinct in Ohio. Publication 5356. Columbus, OH. ODW. 2013. State listed species. Available online at http://www.dnr.state.oh.us/Home/ ExperienceWildlifeSubHomePage/Endangeredthreatenedspeciesplaceholder/resourcesmgtplansspecieslist/ tabid/5664/Default.aspx. Accessed 20 June 2013. Ojeda, R.A., J. Stadler, and R. Brandl. 2003. Diversity of mammals in the tropical-temperate neotropics: Hotspots on a regional scale. Biodiversity Conservation 12:1431–1444. Osbourne, J.D., J.T. Anderson, and A.B. Spurgeon. 2005. Effects of habitat on small-mammal diversity and abundance in West Virginia. Wildlife Society Bulletin 33:814–822. Panzacchi, M., J.D.C.Linnell, C. Melis, M. Odden, J. Odden, L. Gorini, and R. Andersen. 2010. Effect of land-use on small-mammal abundance and diversity in a forest-farmland mosaic landscape in southeastern Norway. Forest Ecology and Management 259:1536– 1545. Poppitt, S.D., J.R. Speakman, and P.A. Racey. 1993. The energetics of reproduction in the Common Shrew (Sorex araneus): A comparison of indirect calorimetry and the doubly labeled water method. Physiological Zoology 66:964–982. Powers, D.M. 2011. Physiographic provinces map, modified from physiographic regions of Ohio, ODNR, Division of Geological Survey, 1998. Page-size map with text, scale 1:2,100,00. Available online at http://www.ohiodnr.com/Home/Nature/Five Physiographic Regions. Accessed 20 May 2011. Reed, A.W., and N.A. Slade. 2008. Density-dependent recruitment in grassland small mammals. Journal of Animal Ecology 77:57–65. Rich, S.M., C.W. Kilpatrick, J.L. Shippee, and K.L. Crowell. 1996. Morphological differentiation and identification of Peromyscus leucopus and P. maniculatus in northeastern North America. Journal of Mammalogy 77:985–991. Rickard, W.H. 1960. The distribution of small mammals in relation to climax vegetation mosaic in eastern Washington and northern Idaho. Ecology 41:99–106. Schipper, J., et al. 2008. The status of the world’s land and marine mammals: Diversity, threat, and knowledge. Science 322:225–230. Schwartz, O.A., and P.D. Whitson. 1987. A 12-year study of vegetation and mammal succession on a reconstructed tallgrass prairie in Iowa. American Midland Naturalist 117:240–249. Speakman, J. 1997. Factors influencing the daily energy expenditure of small mammals. Proceedings of the Nutrition Society 56:1119–1136. Svendson, G.E. 1976. Microsorex hoyi in southeastern Ohio. Ohio Journal of Science 76:102. Vessey, S.H., and K.B. Vessey. 2007. Linking behavior, life histories, and food supply with population dynamics of White-footed Mice (Peromyscus leucopus). Integrative Zoology 2:123–131. Vick, M.H. 2004. Inventory of distribution, composition, and relative abundance of mammals at Hopewell Culture National Historical Park. National Park Service. Available online at http://science.nature.nps.gov/im/units/htln/library/Inventory/HOCU_ Mammals_200412_i.pdf. Accessed 20 July 2013. Whitaker, J.O., Jr. 1967. Habitat relationships of four species of mice in Vigo County, Indiana. Ecology 48:867–872. Northeastern Naturalist Vol. 21, No. 2 J.D. Harder, J.K. Kotheimer, and I.M. Hamilton 2014 233 Whitaker, J.O., Jr., B.L. Walters, and S. Johnson. 2007. Terrestrial mammals of the Naval Support Activity (NSA Crane) in Martin County and of the Crane Division of the Naval Surface Warfare Center (Glendora site) in Sullivan County, Indiana. Proceedings of the Indiana Academy of Science 116:84–89. Williams, D.F., and S.E. Braun. 1983. Comparison of pitfall and conventional traps for sampling small-mammal populations. Journal of Wildlife Management 47:841–845. Wrigley, R.E., J.E. Dubois, and H.W.R. Copland. 1979. Habitat, abundance, and distribution of six species of shrews in Manitoba. Journal of Mammalogy 60:505–520. Yoccoz, N.G., J.D. Nichols, and T. Boulinier. 2001. Monitoring of biological diversity in space and time. Trends in Ecology and Evolution 16:446–453.