e operated for 3 consecutive nights and checked every morning. Identification of all animals investigating the scent station was made by the tracks left in the scent station. Scent stations were monitored in conjunction with small mammal traps during May, June, and July 2008 and 2009.
2.9. Statistical Analysis
Waterbird abundance (number/39.25 ha plot) of combined waterbirds and those species representing 2% or more of detections, species richness, and Shannon Diversity Index (diversity) (dependent variables) were rank transformed  and compared among sites (independent variable) using a repeated measures single-factor analysis of variance (ANOVA) with an alpha level of 0.05. A Tukey’s honest significance difference multiple comparison post hoc test was used for pairwise comparisons of significant variables. Flyovers were excluded from analyses. For the islands with bridges, data on whether the birds flew over or under the bridge also were summarized to investigate response to bridge.
Songbird abundance (individuals/ha) for all species combined and for any species representing 2% or more of observations, species richness, and diversity (dependent variables) were rank transformed  and compared by site, distance, and the interaction term (independent variables) using a repeated measures 2-way ANOVA (α = 0.05). Because the levels of distance had no physical meaning for Muskingum, a partial analysis was conducted to examine distance only for Blennerhassett and Buckley islands. In the case where there was no significant interaction, a contrast for the effect of distance pooled across both islands was estimated and, if necessary, follow up pairwise comparisons between levels of distance were estimated. If there was a significant interaction, then the effect of distance was examined for the 2 islands on an individual basis (i.e., a simple effects analysis). A model allowing for heterogeneous variance was fitted for all variables due to nonconstant variance. Residual diagnostics were then used to verify the assumptions of the model. The test of interaction was done using untransformed data to obtain a more approximate analysis . When subsequently testing main effects or contrasts, rank transformed data were used.
We initially evaluated the uniqueness of species composition by site and distance from bridge (independent variables) with PERMANOVA  using the Vegan package in Program R. A subsequent one at a time analysis was done using multi-response permutation procedures (MRPP) in PC-ORD version 5.10 software . MRPP is a non-parametric multivariate procedure for testing the hypothesis of no differences in species composition of 2 or more a priori groups of plots . For these analyses, all species that occurred less than 3 times were deleted. Individuals/ha (dependent variable) were calculated for abundance data and were relativized using a general column relativization. A Sorensen (Bray-Curtis) distance measure was used with a rank transformed distance matrix and an alpha level of 0.05. A Sorensen community similarity index, which compares pairwise similarities, was used to evaluate songbird overlap among sites and distances.
Lastly, the affinity of species occurrence by site and by distance from bridge were calculated using indicator species analysis . Indicator species analysis combines information on the species abundance at a particular treatment and the faithfulness of occurrence of a species at a particular treatment. First, an indicator value for each species was calculated with a value between 0 and 100. A value of 0 showed no indication (i.e., no association) of the species and a value of 100 showed perfect indication of the species. Second, the statistical significance of the highest indicator value (IVmax) for each species was tested using a Monte Carlo test using 4999 permutations with an alpha level set at 0.05.
Combined anuran abundances, combined call index codes (CIC), abundances and CIC of those species representing 2% or more of all detections, species richness, and diversity (dependent variables) were rank transformed  and compared among sites and distances (independent variables) using a single-factor ANOVA with an alpha level of 0.05. Due to anurans not being detected at Buckley or Muskingum, only Blennerhassett and Grape were used in analyses. Site comparisons for Blennerhassett and Grape used transects only in which anurans were detected. Because Grape did not have sufficient distances, distance comparisons were made only at Blennerhassett.
Turtle relative abundance (captures/100 trap nights) for combined turtles and any species representing 2% or more of all captures, species richness, and diversity (dependent variables) were rank transformed  and compared among sites, distances, and the interaction term (independent variables) using a 2-way ANOVA with an alpha level of 0.05 and proceeded in the same fashion as the songbird analysis with a partial analysis for distance using only Blennerhassett and Buckley islands. Data were combined over each 3 day trapping interval and over the 3 trapping periods for each individual trap giving each trap a potential of 9 trap nights. Each trapping location was counted as a replicate. Recaptures were excluded from abundance analyses.
Small mammal relative abundance (captures/100 trap nights) of combined small mammal captures and individual species representing at least 2% of total captures, species richness, and diversity (dependent variables) were rank transformed  and compared by site, distance, and the interaction term (independent variables) using a repeated measures 2-way ANOVA (α = 0.05) in the same fashion as the songbird and turtle analyses. Abundance data were combined over each 3 night interval as 1 total trapping period. Deer mice (Peromyscus maniculatus) and white-footed mice (Peromyscus leucopus) were combined into a single species (Peromyscus spp.) to account for identification biases. All recaptures were excluded from analyses.
Scent station data for furbearers were analyzed based on presence or absence by G-tests of independence with William’s correction factor. Data were combined for each scent station over the 3 consecutive nights each station was operated and over the 6 survey periods in 2 years. Data were then combined by distance for each site. Analyses were run for all furbearer species combined (whether any furbearer tracks were present) and for each individual species representing 2% or more of furbearer species identified. Analyses compared both site and distance from bridge by combining total presences and absences for each site or distance. Each adjusted G-statistic was tested and an alpha level of 0.05 was set. Pair wise comparisons (among sites or distances) of significant tests were analyzed similarly. Transects at Muskingum Island were not used in the distance comparison due to no bridge being present at that site.
A total of 19 waterbird species (n = 11 at Blennerhassett, n = 8 at Buckley, and n = 17 at Muskingum) was observed during surveys (see  for a complete list). Additionally, 15 other waterbird species were observed within the study area but not during survey hours . Combined waterbird abundance (individuals/39.25 ha plot) was not significantly higher at Blennerhassett compared to Buckley or Muskingum although Blennerhassett had 4 times as many waterbirds as Buckley (Table 1).
Only Canada goose (Branta canadensis) (n = 1007, 73.6%), double-crested cormorant (Phalacrocorax auritus) (n = 28, 2.1%), great blue heron (Ardea herodias) (n = 36, 2.6%), mallard (Anas platyrhynchos) (n = 124, 9.1%), and wood duck (Aix sponsa) (n = 54, 3.9%) had enough observations to analyze separately. Canada goose, double-crested cormorant, and wood duck abundances did not differ significantly among sites (Table 1). Great blue heron abundances were greater at Muskingum than at Blennerhassett or Buckley, which were similar. Mallard abundances were similar between Blennerhassett and Muskingum, but were significantly lower at Buckley.
Overall species richness was significantly lower at Buckley than the other 2 sites (Table 1). Overall diversity differed among all 3 sites with Muskingum having the highest diversity and Buckley having the lowest. At Blennerhassett, 185 of 239 (77.4%) observations of waterbirds flew over the bridge and 54 of 239 (22.6%) flew under the bridge. At Buckley, 138 of 191 (72.3%) observations
Table 1. Summary of waterbird variables (abundances in the form of individuals/39.25 ha plot) with their means and standard errors for Blennerhassett, Buckley, and Muskingum islands, West Virginia, USA for 2007-2009 with same letters indicating no significance among sites (bolded means are significant at α = 0.05).
of waterbirds flew over the bridge and 54 of 191 (27.7%) flew under the bridge.
Songbird surveys detected a total of 60 species (n = 48 at Blennerhassett, n = 53 at Buckley, and n = 42 at Muskingum) . A total of 17 species resulted in enough observations to analyze separately by sites and distances (Table 2). Three of these species had a significant site by distance interaction with Carolina chickadee (Poecile carolinensis) having higher abundances at the 0 m transect at Muskingum than at the 0 m transects at Blennerhassett and Buckley, while house wrens (Troglodytes aedon) had higher abundances at the 0 m transects at Blennerhassett compared to the 0 m transects at Buckley and Muskingum, and rock pigeon (Columba livia) was only found at the 0 m transects at Blennerhassett and Buckley causing an interaction and indicating an overall bridge effect.
Seven species differed significantly by site. Red-winged blackbird (Agelaius phoeniceus) and song sparrow (Melospiza melodia) were higher at Blennerhassett, gray catbird (Dumetella carolinensis) and red-eyed vireo (Vireo olivaceus) were lower at Blennerhassett, Carolina wren (Thryothorus ludovicianus) and tufted titmouse (Baeolophus bicolor) were higher at Muskingum, and European starling (Sturnus vulgaris) and yellow warbler (Dendroica petechia) were higher at Buckley (Table 2). Three species significantly differed by distances. Common yellowthroat (Geothlypis trichas) and Carolina wren were lower at 0 m transects and European starling was higher at 0 m transects (Table 2). No other species significantly differed among sites, distances, or interaction.
All species that occurred less than 3 times were removed from the total species composition matrix reducing the 60 to 46 for the site analysis and 44 for the distance analysis for MRPP . Results of the PERMANOVA showed no evidence of a site by distance interaction (F4, 25 = 0.98, p = 0.53). Results of the MRPP suggest that all 3 islands have different songbird communities and that the 0 m transects have significantly different songbird communities than 100 m and 300 m transects (p ≤ 0.005). The Sorensen community similarity index comparing songbird composition among sites (0.402) and distances (0.456) further indicated different communities (p ≤ 0.016).
Results from indicator species analysis suggest certain species were more likely to occur on particular sites and at particular distances from the bridge . Song sparrow was indicative of Blennerhassett, house finch (Carpodacus mexicanus) was indicative of Buckley, and 12 species were indicative of Muskingum. Two species cliff swallow (Petrochelidon pyrrhonota) and rock pigeon were indicative of 0 m transects.
A total of 7 anuran species were detected at Blennerhassett Island and 5 anuran species at Grape Island: spring peeper (Pseudacris crucifer), Fowler’s toad (Anaxyrus fowleri), eastern American toad (Anaxyrus americanus americanus), American bullfrog (Lithobates catesbeianus), Cope’s gray treefrog (Hyla chrysoscelis), pickerel frog (Lithobates palustris), and northern green frog (Lithobates clamitans melanota). Only observations of spring peeper (n = 674, 88.5%), Cope’s gray treefrog (n = 39, 5.1%), American bullfrog (n = 19, 2.5%), and northern green frog (n = 18, 2.4%) resulted in enough detections to analyze separately.
Combined anuran abundances (individuals/transect) were similar between Blennerhassett and Grape islands and at 0, 100, and 300 m transects (p = 0.400) . Likewise, spring peeper, American bullfrog, Cope’s gray
Table 2. Summary of songbird variables (abundances in the form of individuals/ha) with their means and standard errors for Blennerhassett, Buckley, and Muskingum islands, West Virginia, USA and at 0, 100, and 300 m from the bridge for 2008-2009 with same letters indicating no significance among sites or distances (bolded means are significant at α = 0.05).
treefrog, and northern green frog abundances did not show significant difference between sites or among distances (p ≥ 0.623). Call index codes for combined anurans also were similar between Blennerhassett and Grape and among 0, 100, and 300 m transects at Blennerhassett (p ≥ 0.400). Similarly, spring peeper, American bullfrog, Cope’s gray treefrog, and northern green frog call index codes did not differ between sites or among distances (p ≥ 0.560). Species richness and diversity also were similar between sites and among distances (p ≥ 0.402).
A total of 728 trap nights were attempted with 651 total trap nights occurring after deductions. We had 96 captures of 88 individuals from 5 species: eastern spiny softshell (Apalone spinifera spinifera), common snapping turtle (Chelydra serpentina), northern map turtle (Graptemys geographica), Midland painted turtle (Chrysemys picta marginata), and stinkpot (Sternotherus odoratus). Of these 88 individuals, a total of 11 common snapping turtles and 4 Midland painted turtles were captured in the slough on Blennerhassett Island.
Combined abundances (captures/100 trap nights) were not different among sites, distances from the bridge, or by the interaction term (Table 3). Eastern spiny softshell (n = 46, 52.3%), common snapping turtle (n = 31, 35.2%), northern map turtle (n = 6, 6.8%), and Midland painted turtle (n = 4, 4.5%) occurred in large enough numbers to analyze separately. Eastern spiny softshell, common snapping turtle, and northern map turtle abundances did not differ significantly among sites, distances, or by the interaction term (Table 3). Midland painted turtle abundances significantly differed among sites due to that species only being captured at Blennerhassett (Table 3). Species richness did not differ among sites, distances, or by the interaction term (Table 3). Diversity differed significantly among sites due to Blennerhassett and Grape
Table 3. Summary of turtle variables (abundances in the form of captures/100 trap nights) with their means and standard errors for Blennerhassett, Buckley, Muskingum, and Grape islands, West Virginia, USA and 0, 100, and 300 m from the bridge for 2008 and 2009 with same letters indicating no significance among sites or distances (bolded means are significant at α = 0.05).
differing from Buckley and Muskingum, but did not differ by distance, or by the interaction term (Table 3).
3.5. Small Mammals
A total of 16,448 trap nights were attempted with 14,625 total trap nights calculated after deductions yielding a total of 1124 captures of 733 individuals. A total of 9 small mammal species were encountered during trapping surveys . Peromyscus spp. (n = 495, 67.5%), meadow vole (Microtus pennsylvanicus) (n = 183, 25.0%), and eastern chipmunk (Tamias striatus) (n = 24, 3.3%) occurred in enough observations to analyze separately. Three variables, Peromyscus spp., species richness, and diversity, had a significant interaction effect (Table 4). These 3 variables had a significant interaction effect due to the 0 m transects at Blennerhassett having lower abundances, richness, and diversity under the bridge compared to 100 and 300 m transects. Also, the 100 and 300 m transects at Blennerhassett differed from the 100 and 300 m transects at Buckley and Muskingum due to higher abundances, richness, and diversity at the 100 and 300 m transects at Blennerhassett.
Combined small mammal abundances differed significantly by distances due to lower abundances at the 0 m transects, but did not differ among sites or by the interaction term (Table 4). Meadow vole abundances differed by sites due to this species not being sampled at Muskingum, but did not differ among distances or by the interaction term. Eastern chipmunk abundances did not differ among sites, distances, or by the interaction term.
A total of 4 furbearer species were encountered in scent station surveys: raccoon (Procyon lotor) represented 82.8% of all occurrences, red fox (Vulpes vulpes) 8.2%, Virginia opossum (Didelphis virginiana) 7.5%, and American beaver (Castor canadensis) 1.5%. Occurrences of all furbearer species combined and Virginia opossum did not differ by site or by distance (Table 5). Raccoon occurrences differed by site and by distance with transects at 0 m at Blennerhassett having lower occurrences. Red fox occurrences differed by site with Blennerhassett having the highest occurrences, but not by distance (Table 5). American beaver occurrences were not tested due to too few observations.
Of the 5 species analyzed, only great blue heron and mallard differed among sites. Great blue heron densities were 3.84 times higher at Muskingum than Blennerhassett and 13.67 times higher at Muskingum than Buckley. This could be attributed to great blue herons being negatively impacted by the highway bridge crossings at the other 2 islands. Great blue herons prefer remoteness and
Table 4. Summary of small mammal variables (abundances in the form of captures/100 trap nights) for Blennerhassett, Buckley, and Muskingum islands, West Virginia, USA and 0, 100, and 300 m from the bridge for 2008 and 2009 with same letters indicating no significance among sites or distances (bolded means are significant at α = 0.05).
Table 5. Summary of scent station data for furbearers with their presence (P) and absence (A) percentages (%) for Blennerhassett, Buckley, and Muskingum islands, West Virginia, USA and 0, 100, and 300 m from the bridge for 2008 and 2009 with the same letters indicating no significance among sites or distances (bolded species are significant at α = 0.05).
lack of human disturbance [47,48]. Noise can disrupt heron breeding and nesting patterns . Additionally, great blue heron nesting patterns and nest success can be altered by other human disturbances including: increased human presence, land development, and destruction of habitat . Mallard abundance was lower at Buckley Island possibly indicating that over time bridges impact mallards, but because mallard abundances were similar at Blennerhassett and Muskingum, we believe the Blennerhassett Island Bridge has no current impacts to mallards. There was no evidence that the Blennerhassett Island Bridge caused negative impacts to abundances of any other waterbird species.
We found waterbird species richness and diversity were lower at Buckley compared to the other 2 islands. This may be the result of the bridge at Buckley, but this is difficult to determine due to low abundances of 6 species found only at Muskingum. Lower species richness and diversity at Buckley are more easily explained by natural variation in observations rather than from the presence of a bridge.
Waterbirds do collide with bridges, which create barriers to movement and causes mortality and population declines. Our data, however, shows no indication that the bridge was a barrier or cause of mortality to waterbirds during diurnal time periods. Waterbirds more often flew over the bridge as opposed to under the bridge; however, waterbirds did not hesitate to fly under the bridge when flying low over the water. We had direct observations of many species including: belted kingfisher (Megaceryle alcyon), Canada goose, double-crested cormorant, great blue heron, mallard, spotted sandpiper (Actitis macularia), and wood duck flying under the bridge. The bridge does not seem to be a major barrier for waterbirds during flight, likely due to its relatively low height (24 m above the surface). There were no observations or documentations of waterbirds colliding with the bridge during diurnal flight.
Our study indicates that the Blennerhassett Island Bridge has some, albeit minor impacts to songbirds that resulted in a decrease in overall songbird abundance and changes in species composition most likely due to the area directly under the bridge being highly disturbed and lacking woody vegetation. This effect was mostly due to the removal of all trees, snags, and coarse woody debris and may be responsible for decreases in forest interior songbirds. Rock pigeon, cliff swallow, and European starling had high affinity for transects under bridges as these are all generalist species that use bridges and other manmade structures for nesting, roosting, and perching . House wrens had high abundances under the Blennerhassett Island Bridge due to the placement of songbird nest boxes under the bridge after construction.
Songbird species composition is closely related to habitat structure [51,52]. Our study echoes this as songbirds observed were closely related to known habitat preferences with habitat generalist species using the areas directly under the bridges. Muskingum Island had higher species richness and diversity, different composition, and higher abundances of forest interior songbirds than the other 2 sites. This is largely due to all transects at Muskingum Island being located in bottomland hardwood forest  which harbor high species richness of passerines  while transects at Blennerhassett and Buckley islands varied across cover types with transects in old-field habitat having lower species richness.
Changes in songbird communities found in our study follows previously documented findings on the changes in species composition following a disturbance. Songbird communities vary among successional stages and forest types, and a diversity of landscapes maintains songbird diversity. Community changes of songbirds similar to those under the bridge have been documented for many types of natural and anthropogenic disturbances includeing: fire, blow downs, timber harvests, mowing, agriculture, residential development, and urbanization [51,54- 57].
Anurans are commonly depicted as ecologically sensitive to anthropogenic disturbance [58,59]. However, we found anuran abundances, richness, and diversity were not different between a site with a large bridge crossing and a site without a bridge. Also, anuran abundances, richness, and diversity were no different at the different distances at Blennerhassett Island. Species were found calling throughout the wetland complex on the island and also in the wetland vegetation under the bridge on the West Virginia mainland. This indicates that the bridge had no negative effect on anuran relative abundance. Calling indices did not differ between sites or among distances further indicating the Blennerhassett Island Bridge has no negative effect on abundance or richness.
All anuran species encountered in this study had been previously documented on the Ohio River Islands National Wildlife Refuge and are common along the Ohio River . Habitat was the determining factor of anuran presence , and anuran abundance and richness in relation to the bridge indicated the bridge had no negative effect. Other research has indicated that anurans respond positively or neutrally to disturbance. Two endangered anuran species responded favorably or neutrally to ground disturbance on a military training area in Germany . Populations of anurans in floodplain wetlands were similar in abundance and richness compared to less dynamic wetlands not in floodplains . Also, anuran abundance and richness was stable at a suburban wildlife refuge despite large amounts of development in close proximity to wetlands .
Turtles are tolerant of urbanization and anthropogenic changes in the landscape [65,66]. Our study indicated that turtle abundance and richness were no different among sites or at the different distances from the bridges indicating that the new bridge has no negative effect on these metrics. Additionally, the difference in diversity among sites had no attribution to the new bridge. Eastern spiny softshell and common snapping turtle represented the majority of captures which is comparable to other studies on the Ohio River . Eastern spiny softshells were encountered only in the main river which coincides with habitat preferences for this species . Common snapping turtles were found readily in the main river at each of the study islands but more frequently in back channels . We believe the Ohio River supports a large population of these 2 species around the study areas due to quality and quantity of habitat in the form of sand and gravel bars, emergent vegetation, and woody debris.
The other 3 species (northern map turtle, Midland painted turtle, and stinkpot) were all encountered in low numbers. Each species is associated with slow bodies of water with soft bottoms and abundant basking sites  which are plentiful around Ohio River islands and in the slough on Blennerhassett, and we do not believe the bridge is affecting their abundances. Aquatic turtles are disturbance tolerant animals unlike many other taxa of wildlife. Spiny softshell turtle survival in Arkansas was not negatively affected when creek beds were bulldozed and woody debris was removed which decreased habitat quality and quantity . Painted turtle nesting ecology showed no changes with increases of human recreation at a major nesting beach . Species of freshwater turtles, including those found in this study, persist following disturbances in the form of urbanization  and in some cases may even be more abundant in urban habitats than they are in more undisturbed natural areas [69,70].
4.5. Small Mammals
Small mammals are typically depicted as disturbance tolerant [71,72]. However, we found that combined small mammal abundance and abundances of Peromyscus spp. were lower under the new bridge indicating that it has impacts. Species richness and diversity also were lower under the new bridge compared to other distances indicating that the disturbance from construction of the bridge may result in a decrease in small mammal richness and diversity. We believe these impacts are due to the amount and type of vegetative cover, and preferences of microhabitat by certain species . We suggest these impacts are temporary and that abundance, richness, and diversity under the bridge will increase to levels similar to other distances as shown at the older Buckley Island Bridge making it consistent with other studies following a disturbance [74,75].
Total abundances of small mammals were highest at Blennerhassett, which is most likely due to more quality, quantity, and diversity of habitat. Lack of grasslands, old fields, and brushy areas at Muskingum Island  may be causing lower abundances , while possible reasons for lower abundance at Buckley Island include high numbers of raccoons preying on small mammals  or high intensity disturbance from agriculture, residential development, and urbanization. Also, all 3 sites may be experiencing variation due to cyclical fluctuations of small mammal populations [78,79].
The vegetation under the new bridge was of low coverage and consisted of herbaceous, disturbance tolerant plants typically used for reclamation making it similar to roadsides and may be the reason why abundances, richness, and diversity was low. However, capture rates and diversity increased under the bridge as vegetative cover increased during the study. We hypothesize that the low abundances, richness, and diversity under the bridge is temporary and will increase as vegetative coverage increases. Also, there was only one detection of a small mammal moving from transect to transect at all of the sites (one Peromyscus spp. moved from a 100 m transect to a 0 m transect at Blennerhassett) possibly indicating that small mammal populations were impacted under the bridge during construction and are slowly increasing as habitat is restored.
We found that combined furbearer detections did not differ among sites or distances. However, 2 species, raccoon and red fox, did vary. Raccoon detections were much lower at 0 m transects at Blennerhassett. These impacts could possibly be from habitat fragmentation and loss, most notably disturbance of natural shorelines caused by construction of the bridge. However, increased urbanization and habitat modification and fragmentation result in higher raccoon densities . Raccoons respond well and use wildlife passages and underpasses  and we assume they use the area under the bridge similarly. We hypothesize that raccoons commonly frequent the area beneath the bridge, but may be patrolling around the wetland instead of the shorelines, thus avoiding our scent stations. Raccoon populations within the study area appear to be high due to high presence shown in scent station surveys, large amounts of tampering of small mammal traps during small mammal trapping periods, abundant quality and quantity of habitat (proximity to water, proximity to urban areas, and multiple cover types), lack of hunting and minimal trapping on the islands, and abundant food sources [77,81].
Red fox occurrences were highest at Blennerhassett and this species did not occur at Muskingum which is the reason for variation among sites and indicates the new bridge has no negative impacts. Personal observations suggest red fox populations are high on Blennerhassett as this island has preferred habitats including: forest, open areas, and grasslands and mixed vegetation communities such as edge habitats and mixed scrub and woodland .
Any disturbance is likely to have impacts on some species, either negatively or positively. Species adapted to man-made structures such as rock pigeons responded favorably to the bridge and some such as Carolina wrens responded negatively. However, most songbird and waterbird species displayed no effect. Indeed, variation among islands was greater than variation among distances for almost all taxa. The more aquatic taxa such as turtles and anurans showed no relative density effects from the bridge. Overall, the new bridge had little significant impact on relative density of wildlife, but the bridge did impact community composition and relative abundance of some species. It is possible that more impacts would have been seen if the island did not have a long history of human impacts. Continued monitoring of this and other bridges is necessary to determine potential long-term impacts.
We thank the West Virginia Department of Transportation, Division of Highways for the funding of this project. We thank the US Fish and Wildlife Service, the West Virginia Division of Natural Resources Wildlife Resources Section and Recreation and Parks Section for their cooperation throughout this project. We thank Dr. Phil Turk, Dr. George Merovich and Dr. Jim Rentch for their assistance with statistical analysis for this project. Special thanks is also extended to the many biologists, field technicians, graduate students, and volunteers who assisted with this project including: Patty Morrison, Gary Batton, Jeff McCrady, Matt Baker, Craig Leibfreid, Graham Nesselrodt, Matt Coen, Brandon Miller, Tristan Gingerich, Walter Veselka, Gabe Strain, Kyle Aldinger, Mandy Weston, Curtis DeLong, Brian Reed, and David Vance. This is scientific article number 3143 of the West Virginia University Agriculture and Forestry Experiment Station.
Appendix 1. Common and Scientific Names of Species in Tables