ENVIRONMENTAL FACTORS STRUCTURING TERRESTRIAL HERPETOFAUNAL COMMUNITIES AT THE ARKANSAS POST NATIONAL MEMORIAL, ARKANSAS

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Forest Resources

by

Kevin Ryan Rose, B.S. Auburn University, 2004

August 2007 University of Arkansas - Monticello

ABSTRACT Worldwide amphibian and reptile population declines have highlighted the need for management of these species. However, little is known about environmental factors that structure herpetofaunal community composition. I conducted this study to determine the environmental factors that structure amphibian and reptile communities at Arkansas Post National Memorial, Arkansas (the Post). Herpetofaunal communities were surveyed using area constrained searches and environmental factors were recorded. Data were analyzed using multivariate analysis. I recorded 2934 individuals representing 23 species. Species were not evenly distributed across the Post. My results indicated that moisture was the main environmental factor structuring amphibian and reptile communities at the Post. Amphibians were more closely associated with soil moisture than reptiles. Reptiles were more influenced by presence of snags and standing water. These results support the findings of previous studies. These studies concluded that environmental factors such as canopy cover, litter depth, woody plant cover, and large down wood that affect moisture levels were important to amphibians and reptiles. My data can be used by the National Park Service to make more informed management decisions. I recommend that forest-to-water edge habitat be protected as these areas have the highest species richness.

iii

ACKNOWLEDGEMENTS

I would like to acknowledge those who have contributed to this project. First, I would like to thank my family for their support as I find my way through life. My wife, Jacqueline Rose, has been my companion, my rock, and a valuable member of my field crew. My parents, David and Mary Ann Rose, have provided encouragement and have enabled me to pursue my dreams. My brother, David “Chip” Rose, who agreed to work for me, made the extended field work possible and enjoyable. I would like to thank my advisor, Dr. Don White, for this opportunity and his guidance. I would also like to thank him for the freedom that he allowed me to independently work through this project while still providing support when needed. I thank my committee members, Dr. Robert Kissell and Dr. Philip Tappe for their advice and insights on this project. Chris Watt was instrumental in the preparation, setup, and execution of data collection and analysis. I also thank Leo Acosta for his contribution to data collection. Finally, I would like to thank the Arkansas Forest Resources Center and the National Park Service for funding this project which has allowed me to attend graduate school.

iv

TABLE OF CONTENTS Page ABSTRACT....................................................................................................................... iii ACKNOWLEDGEMENTS............................................................................................... iv TABLE OF CONTENTS.................................................................................................... v LIST OF TABLES............................................................................................................. vi LIST OF FIGURES .......................................................................................................... vii INTRODUCTION .............................................................................................................. 1 OBJECTIVE ....................................................................................................................... 3 LITERATURE REVIEW ................................................................................................... 4 Community Ecology ....................................................................................................... 4 Evolution....................................................................................................................... 11 Habitat Fragmentation .................................................................................................. 11 Environmental Factors .................................................................................................. 13 Gradient Analysis.......................................................................................................... 17 Ordination Methods .................................................................................................. 17 Description................................................................................................................ 18 Ordination Diagrams................................................................................................ 20 METHODS ....................................................................................................................... 23 Study Site ...................................................................................................................... 23 Vegetation Types .......................................................................................................... 27 Plot Establishment ........................................................................................................ 30 Amphibian and Reptile Surveys ................................................................................... 32 Habitat Measurements .................................................................................................. 34 Diversity Analysis......................................................................................................... 39 Gradient Analysis.......................................................................................................... 39 RESULTS ......................................................................................................................... 43 Survey Results .............................................................................................................. 43 Principal Components Analysis.................................................................................... 49 Partial Redundancy Analysis ........................................................................................ 53 DISCUSSION ................................................................................................................... 65 Surveys.......................................................................................................................... 65 Species Richness........................................................................................................... 66 Diversity Indices ........................................................................................................... 66 Communities ................................................................................................................. 66 Management Recommendations................................................................................... 70 LITERATURE CITED ..................................................................................................... 71

v

LIST OF TABLES Table

Page

1. Climatic conditions for Dumas, Arkansas 2006 during sampling periods…………………………………………………………………………....26 2. Vegetation type, area, and number of plots used to sample amphibians, reptiles, and habitat characteristics at the Arkansas Post National Memorial, 2005-2006……………………………………………………………31 3. Habitat variables sampled within each sampling plot at the Arkansas Post National Memorial, 2005-2006. Location refers to the location in the plot the variable was measured……………………………………………………….35 4. Daubenmire cover-class scale used to estimate percentage cover at the Arkansas Post National Memorial, 2005-2006....………………………………..38 5. Gradient lengths from detrended analyses expressed as standard deviations. If gradient lengths were <3.0 standard deviations, linear methods were chosen for data analysis………………………………………………………….42 6.

Species codes, common names, scientific names, number recorded, and proportion of total sample………………………………………….…………….44

7. Mean species richness (S), Shannon-Weaver diversity indices (H), and Shannon-Weaver evenness indices (J) for each sampling plot on the Arkansas Post National Memorial, 2005-2006 across all sampling periods…….46 8. Environmental factor-to-axes correlations and P-values for Figure 14………….55 9. RDA summary for Figure 14…………………………………………………….56 10. Environmental factor-to-axes correlations and P-values for Figure 15………….59 11. RDA summary for Figure 15…………………………………………………….60 12. Environmental factor-to-axes correlations and P-values for Figure 16………….63 13. RDA summary for Figure 16…………………………………………………….64

vi

LIST OF FIGURES Figure

Page

1. Clements’s concept of community organization. Lines represent population abundance of individual species along an environmental gradient……………………………………………………………………………6 2. Gleason’s concept of community organization. Lines represent population abundance of individual species along an environmental gradient……….………9 3. RDA ordination diagram displaying samples (circles), species (triangles), and environmental variables (arrows)……………………………………………21 4. Location of the Arkansas Post National Memorial……………………………....24 5. Vegetation types and sample plot locations at the Arkansas Post National Memorial, 2005-2006……………………………………………………………25 6. Plot configuration used to sample environmental variables and amphibian and reptile presence and abundance at the Arkansas Post National Memorial, 2005-2006……………………………………………………………33 7. Log decay classes used to classify down logs at the Arkansas Post National Memorial, 2005-2006…………………………………...…………………….…37 8. Total abundance (log transformed) of amphibians and reptiles recorded by sampling plots on the Arkansas Post National Memorial, 2005-2006…………...45 9. Mean herpetofaunal species richness on the Arkansas Post National Memorial, 2005-2006.…………………………………….……………………..47 10. Mean herpetofaunal Shannon-Weaver diversity indices plotted by plot for The Arkansas Post National Memorial, 2005-2006……………….……….…….48 11. PCA ordination diagram of plots sampled at the Arkansas Post National Memorial, 2005-2006……………………………………………………………50 12. PCA ordination diagram of plots sampled at the Arkansas Post National Memorial, 2005-2006. Axes 1 and 2 are plotted with environmental factors projected………………………………………………………………….51 13. PCA ordination diagram of plots sampled at the Arkansas Post National Memorial, 2005-2006. Axes 1 and 3 are plotted with environmental factors projected………………………………………………………………….52

vii

LIST OF FIGURES (continued) Figure

Page

14. RDA ordination diagram of species and environmental factors from sampled plots at the Arkansas Post National Memorial, 2005-2006. Environmental factor-to-axes correlations and P-values are in Table 8. Environmental factor codes are in Table 3 Species codes are in Table 6.…………..…………………..……….……………………………….....54 15. RDA ordination diagram of reptile-environment relationships at the Arkansas Post National Memorial, 2005-2006………………………..…………58 16. RDA ordination diagram of amphibian-environment relationships at the Arkansas Post National Memorial, 2005-2006…………………………………..62

viii

INTRODUCTION Amphibians and reptiles are increasingly the subjects of scientific research, in part due to worldwide herpetofaunal population declines (Barinaga 1990, Blaustein and Wake 1995, Pechmann et al. 1991, Wake 1991, Blaustein et al. 1994, Pounds and Crump 1994, Mount 1996, Gibbons et al. 2000). Losses of amphibian and reptile species may have dramatic affects on entire ecosystems (Whiles et al. 2006). Environmental factors that influence amphibian and reptile community structure are largely unknown due to the wide variation in habitat preferences, modes of locomotion, and life histories among species (Mount 1996, Zug et al. 2001, Trauth et al. 2004). Studies have been conducted on relationships between environmental factors and herpetofaunal community structure. Studies of herpetofaunal communities in Arkansas have been conducted in the Ouachita Mountains (Crosswhite et al. 2004, Shipman et al. 2004, Loehle et al. 2005). Several habitat factors have been suggested to significantly structure amphibian and reptile communities. They include tree basal area, canopy cover (Block and Morrison 1998, Loehle et al. 2005), litter depth, woody plant percent cover, down wood (Dupuis et al 1995, Block and Morrison 1998, Crosswhite et al. 2004), pond hydroperiod (Skelly et al. 1999), salinity of aquatic habitats, slope, temperature, moisture levels, pH (Dunson and Travis 1991), elevation (Hofer et al. 2000), edges between forest and wetlands (Knutson et al. 1999), forest age (Herbeck and Larsen 1999), dominate tree species (Bennett et al. 1980, Hanlin et al. 2000), and habitat fragmentation (Marsh and Trenham 2001, Bell and Donnelly 2006). McCallum et al. (2003) conducted an inventory of amphibians and reptiles at the Arkansas Post National Memorial. They

1

identified 37 species, including 10 anurans, two salamanders, 11 snakes, five lizards, eight turtles, and one crocodilian. Understanding environmental factors associated with herpetofaunal community structure is a fundamental issue in ecological research. As an important part of adaptive natural resource management planning, land managers need information related to factors that structure herpetofaunal communities (Matlack 1994). The National Parks Omnibus Management Act of 1998 requires scientific research and the collection of baseline ecological data from all National Park Service Units. Environmental factors structuring amphibian and reptile communities at Arkansas Post National Memorial have not been examined.

2

OBJECTIVE The objective of this study was to determine the environmental factors that significantly affect the structure of amphibian and reptile communities on the Arkansas Post National Memorial in Arkansas County, Arkansas. I hypothesized that soil moisture would be an important environmental factor structuring amphibian communities, but not as important in reptile communities.

3

LITERATURE REVIEW Community Ecology A biological community is any combination of ecologically-related species that exist together in an area or habitat (Dice 1968, Odum 1971). This classic definition is intentionally broad enough to include groups at multiple spatial scales. Biological communities are divided into major communities and minor communities. Major communities are independent of their surroundings, and therefore, are relatively large and complex. Species within a major community do not interact with species in a different major community. Thus, major communities are closed systems. Minor communities allow for smaller groups of species to be considered communities. Minor communities are not independent of their surroundings and require connections to other minor communities. Immigration and emigration of individuals or energy flow occurs between minor communities (Odum 1971). A minor community can be as small as two organisms of differing species interacting. A major community could be as large as the universe (Dice 1968). Biological communities are typically described by species composition, habitat composition, guilds present, and geographic location. Species composition is used to describe community structure because membership is the defining character of a biological community. However, a simple list of species is insufficient; the abundance of each species must also be determined (Dice 1968). Here, I define herpetofaunal community structure as the occurrence and relative abundance of amphibian and reptile species within a specified area.

4

Ecologists have developed several ideas to explain the complex organization and interactions that constitute biological communities. Diversity indices were developed to describe and compare communities. Many indices use species richness and how evenly the species are represented as measures of diversity. Species richness is simply a count of the number of species within a particular taxa represented in the community. The importance of a species in the diversity index is relative to its abundance. The two most popular diversity indices are Simpson’s index and the Shannon-Weaver index, however there are many others available (Ricklefs 1996, Ricklefs and Miller 1999). While no single concept has explained the structure of all communities, combining the following concepts has allowed for an understanding of the factors that structure communities. Clements (1916) is often credited for developing the concept of plant succession, although it was originally Cowles (1899) that envisioned it. Clements (1916) thought species within a community are closely dependent upon one another and, therefore, the distribution of each species is closely associated to the distribution of the community. As one community replaces another community over time and space, species within a community would eventually be replaced by other species. The lines in Figure 1 represent individual species abundances as they relate to an environmental gradient, say, moisture. Species are grouped into distinct communities, which occur only within lower and upper boundaries of environmental factors. Thus, community distributions result from distinct changes in environment gradients called ecotones (Figure 1; Watt 1964, Ricklefs 1996).

5

Figure 1. Clements’s concept of community organization. Lines represent population abundance of individual species along an environmental gradient. Redrawn from Ricklefs (1996).

6

Gleason (1926), on the other hand, claimed that a biological community is the result of immigration and the ability of the species to tolerate existing environmental conditions. Since immigration and the environment are always changing, the resulting community is unpredictable. A community merely reflects the dynamics of populations as environmental factors and immigration rates change over time and space (Watt 1964). Gleason’s concept of community structure allows the distribution of species to occur independently of other species and allows them to occur in >1 community. Interestingly, species abundance is dependent on environmental variation, but environmental variation is not the only structuring factor (Ricklefs 1996). With increases in some environmental factor, say, moisture, species abundance may increase until it reaches an optimum gain from that environmental factor. As the environmental factor continues to increase, the species decreases until it can no longer withstand the high amounts of this factor. Each species responds to each environmental gradient uniquely (Figure 2). Other succession theories have been proposed to explain changes in community over time. Tansley (1935) emphasized the role of disturbance in structuring communities. These disturbances could be catastrophic, such as a volcanic eruption, or less intense, such as periodic fire. Each disturbance has an effect on the succession of the community. Whittaker (1953) emphasized the affect of the environment, disturbance, and chance on community succession. He claimed that there is not a predetermined end point of succession for any area due to the multitude of factors that affect succession. Westoby et al. (1989) proposed the “state and transition” model. In this concept, states are stable communities that tend not to change over time. Transitions are changes to the states that are initiated by natural or anthropogenic disturbances. By this model,

7

we would expect that communities would not change over time unless there is a disturbance. The intensity of the disturbance influences the next steady state attained. This results in a system that does not change automatically from one community to the next in a predefined order as is predicted by Clementsian succession. Intense disturbance can create a community that Clementsian succession would classify as early successional while moderate disturbance could trigger the transition to the next successional state. The “state and transition” model has gained acceptance in range management recently (Bestelmeyer et al. 2003).

8

Figure 2. Gleason’s concept of community organization. Lines represent population abundance of individual species along an environmental gradient. Redrawn from Ricklefs (1996).

9

Von Liebig (1840) proposed that organisms have requirements for selected factors and the absence or limited supply of one factor will inhibit growth and reproduction. This idea, which has become known as Liebig’s law of the minimum, has been modified to include interactions between factors. For example, high concentrations of one factor, such as sunlight, may increase the need for another factor, such as moisture (Odum 1971). Shelford (1931) increased the scope of Liebig’s law when he proposed that organisms are also limited by maximums. An excessive amount of an essential factor can be as limiting as too little of that factor (Odum 1971). This concept has become known as Shelford’s law of tolerance. Moreover, interactions between factors also occur and influence how the abundance or limitation of one factor affects the increased or decreased tolerance for another factor (Odum 1971). From Liebig and Shelford’s concepts of limiting factors it follows that organisms or communities require complex conditions. However, there are non-environmental factors such as the interactions between species that limit a population or community. Limiting factors tend to be the main factors that structure a community by limiting the abundance of species. Ecologists use this idea to allow them to concentrate on these limiting factors that significantly affect structure of communities. In this way, the effect of every single environmental factor does not need to be accounted for since only a few are the limiting factors in a community (Odum 1971). The concepts proposed by Clements, Gleason, Tansley, Whittaker, Liebig, and Shelford form the theoretical basis for ecological studies along gradients. Whittaker (1967) first popularized gradient analysis in America. Whittaker studied the distribution

10

and abundance of trees along a moisture gradient in Tennessee. He found that each tree species reached peak abundance at different moisture levels. As the moisture increased along the gradient, tree species abundance and composition changed. Each species responded to the moisture gradient differently, a response illustrated by Gleason’s community concept (Figure 2). Evolution Evolution is also an important factor in understanding community structure (Losos et al. 1997, Vitt et al. 2003). Current community composition is dependent on the evolution and relationships of member species just as the evolution of a species is dependent upon the community in which it evolved. Predator-prey relationships and competition are processes of natural selection. These factors affect evolution; therefore evolution is dependent on the community. As a species changes through evolution, its interactions with other species change thereby changing the community. Therefore, communities are structured partially by evolution (Bradshaw and Mortimer 1986). Effects of evolution on community structure are difficult to observe and quantify and may best be studied in communities currently undergoing dramatic changes. Examples may include communities affected by invasive species and habitats altered by humans (Bradshaw and Mortimer 1986). Habitat Fragmentation Habitat fragmentation may be an important factor in structuring amphibian and reptile communities (Marsh and Trenham 2001, Bell and Donnelly 2006) and may be a cause of amphibian population decline (Cushman 2006). Fragmentation occurs when an

11

area is divided into more than one patch and the total area available to species is decreased. This also results in a higher edge-to-area ratio. Metapopulations, partially isolated subpopulations, may be formed by habitat fragmentation. Subpopulations of amphibians are affected by immigration and emigration by individuals from other subpopulations. Subpopulations may go extinct and be recolonized by individuals from other subpopulations. The size of and distance between suitable fragmented habitats affect this interaction. Larger areas tend to support more individuals from more species that smaller areas. Therefore, species in larger areas are less likely to go extinct. Dispersal from one area to another is limited by the distance between areas. Therefore, subpopulations in fragmented habitats that are close together will tend to have more interaction than subpopulations in distant areas (Cox 1997). Quality management practices may require an understanding of metapopulation structure and function (Marsh and Trenham 2001). Bell and Donnelly (2006) studied frogs and lizards in fragmented habitat in Costa Rica. They found frog populations were decreased by forest fragmentation while lizards appeared to be more prevalent in fragmented habitat. They also found the amphibian and reptile community structure in fragmented habitats were different than those in intact forests. Approximately 25% of the species recorded in the continuous forest were not detected in the fragmented forest habitats (Bell and Donnelly 2006). Fragmentation may occur in areas previously thought to be continuous habitat. Roads and abandoned logging roads are common disturbances that may limit suitable habitat for amphibians and reptiles and therefore fragment habitat that appears to be continuous. Semlitsch et al. (2007) found salamander abundance was significantly

12

reduced within 35 m of roadsides and abandoned logging roads in North Carolina. They concluded that areas previously thought of as continuous habitat may actually be divided into small areas of suitable habitat by anthropogenic disturbances. Amphibians and reptile communities may be structured by the effects of habitat fragmentation. The influence may differ among species. The habitat may be fragmented differently for different species depending on scale and habitat requirements. The factors that fragment salamander habitat may not influence a large snake’s habitat. Environmental Factors Several environmental factors have been suggested to significantly structure amphibian and reptile communities. They include: basal area, canopy cover (Block and Morrison 1998, Loehle et al. 2005), litter depth, woody plant cover, down wood (Dupuis et al. 1995, Block and Morrison 1998, Crosswhite et al. 2004), pond hydroperiod (Skelly et al. 1999), salinity of aquatic habitats, slope, temperature, moisture levels, pH (Dunson and Travis 1991), elevation (Hofer et al. 2000), edges between forest and wetlands (Knutson et al. 1999), forest age (Herbeck and Larsen 1999), and dominate tree species (Bennett et al. 1980, Hanlin et al. 2000). Crosswhite et al. (2004) identified canopy cover, litter depth, woody plant cover, and large woody debris as variables that most affected the herpetofaunal community structure in the Ouachita Mountains, Arkansas. They concluded that microhabitat preferences of species explains the differences in herpetofaunal community structure and that reptiles and amphibians respond predictably to changes in habitat structure.

13

Dupuis et al. (1995) found a weak positive relationship between salamanders and understory and down woody debris. They concluded that these environmental factors increase moisture, which explained the response of salamanders. Skelly et al. (1999) found that canopy cover has a negative impact on amphibian larval diversity in ponds. Ponds that have a closed canopy may have lower water temperatures and less vegetation and, therefore, lower dissolved oxygen resulting in less favorable habitat for larval amphibians. They also concluded that hydroperiod controls the number of possible predators as well as the amount of time for larval development. Permanent ponds allow an accumulation of predatory species while temporary ponds support fewer predators. Temporary ponds, however, may not allow enough time for larval development; therefore, ponds with intermediate hydroperiods may be preferred. Loehle et al. (2005) found that young forest stands in Arkansas supported more amphibians and reptiles than older stands, but that reptiles were also abundant in old stands. Reptiles were associated with low basal areas where sun light penetration to the forest floor was high. Amphibians were associated with high basal areas, which probably resulted in higher moisture due to low sunlight penetration. DeGraaf and Rudis (1990) found that forest stands without permanent water supported herpetofaunal communities as rich and diverse as stands with water in New Hampshire. They concluded that balsam fir (Abies balsamea) forests were not as suitable as northern hardwood stands for amphibians after recording half the capture success in conifer forests than in hardwood forests. They hypothesized that the low pH of balsam fir stands accounted for this difference. Redback salamanders (Plethodon cinereus) and red-spotted newts (Notophthalmus viridescens) were reported to be positively associated

14

with understory vegetation and leaf litter depth, which were also associated with hardwood forests. Pough et al. (1987) studied woodland salamander populations in New York and also concluded that populations of woodland salamanders were correlated with understory vegetation and litter depth. Knutson et al (1999) investigated the relationships between anurans and landscape-level environmental factors in Iowa and Wisconsin. They reported that anurans are negatively correlated with urban areas and positively correlated with forests and wetlands. Landscape patterns with high levels of wetland-to-forest edge were the most significant factor in Iowa while forest and agriculture areas were important in Wisconsin. They hypothesized that small patches of forest act as a refuge for anurans in agricultural areas of Wisconsin. They concluded that anurans benefit from edge habitats. Aubry (2000) studied amphibian communities in managed forests in Washington. He found that amphibian populations in managed forests were structured differently than communities in unmanaged forests. Managed forest communities were dominated by three species: northwestern salamander (Ambystoma gracile), western redback salamander (Plethodon vehiculum), and ensatina (Ensatina eschscoltzii) ; communities in unmanaged forests were dominated by six species: northwestern salamander (Ambystoma gracile), western redback salamander (Plethodon vehiculum), ensatina (Ensatina eschscoltzii), tailed frog (Ascaphus truei), roughskin newt (Taricha granulosa), and redlegged frog (Rana aurora). Aubry (2000) concluded that forest age was the greatest factor affecting amphibian community structure. Bennett et al. (1980) compared amphibian communities between three forest types around a lake in South Carolina. They reported that the diversity and abundance of

15

amphibians was higher for a hardwood area than for two pine (Pinus taeda and Pinus elliotti) areas during 1977 but not during 1978. They concluded that the changes in abundance from year to year may be explained by a drought in 1977, which may have allowed the pine areas to become too dry to be suitable for amphibians while the hardwood areas retained moisture longer functioning as a refuge during the drought. Hanlin et al. (2000) studied amphibian activity, abundance, and species richness from 1993 to 1996 at the same study site in South Carolina as Bennett et al. (1980). Since the study by Bennett et al. (1980), the lake was drained and dredged to remove pollution and restore the area to near natural conditions. Hanlin et al. (2000) found that the hardwood forest had the largest abundance of individuals. The slash pine forest (P. elliotti) had the highest evenness and diversity index value. The loblolly pine forest (P. taeda) had the highest species richness. McLeod and Gates (1998) investigated the affects of timber harvest and prescribed burns on amphibian and reptile communities in Maryland. Their study included hardwood, cut-over hardwood, mixed pine-hardwood, and burned pine forests. They found that amphibians were more abundant in hardwood areas than the other areas sampled and that reptiles were more abundant in pine and burned areas. They also found that anurans were least abundant in burned areas. They concluded that forest management practices affected amphibian and reptile communities by affecting moisture levels and sunlight penetration. Herbeck and Larsen (1999) investigated the affects of silvicultural practices on salamander populations in Missouri. They found that young forests supported few salamanders and that salamander abundance increased as the age of the forest increased.

16

They concluded that salamander abundance was affected by the structural differences between forest stand ages. Older forests had more cover objects, leaf litter, and closed canopies, which provided moist microhabitats. Block and Morrison (1998) studied habitat relationships of amphibians and reptiles in California. They found that salamanders were more abundant on north-facing slopes and lizards were more abundant on xeric sites open to sunlight. They concluded that microhabitat preferences at different spatial scales were responsible for species distribution. Hofer et al. (2000) used direct gradient analysis to investigate the relationship between amphibian and reptile communities and elevation, forest type, and water on Mount Kupe, Cameroon. They reported that amphibians responded to the presence or absence of water but reptiles did not. The relationship between amphibians and water disappeared when amphibians were separated by dependence on water for reproduction. Amphibian dependence on water for reproduction was negatively related to elevation since there was less water available to form streams at higher elevations. Gradient Analysis Ordination Methods Gradient analysis can be used to examine variation in community composition by comparing species data to environmental gradients. These gradients can be hypothetical (unconstrained) or measured (constrained). Ordination, a type of gradient analysis, attempts to define axes that account for variability between samples. Unconstrained ordination methods, such as principal components analysis (PCA) and correspondence analysis (CA), define axes based solely on the species composition of the samples. The 17

axes created may or may not represent a definable gradient in the environment. Constrained ordination methods, such as redundancy analysis (RDA) and canonical correspondence analysis (CCA), define axes that are linear combinations of the measured environmental factors. These axes only account for variation in the species data that can be attributed to measured environmental factors (ter Braak 1986, Palmer 1993, ter Braak and Šmilauer 2002, Lepš and Šmilauer 2003). These ordination methods also have partial versions. Partial versions of ordination methods remove the variability in species data associated with covariables before defining axes. Ordination methods can use a linear model or a unimodal model (ter Braak and Šmilauer 2002, Lepš and Šmilauer 2003). Description ter Braak and Šmilauer (2002) and Lepš and Šmilauer (2003) provide a simplified explanation of PCA, CA, CCA and RDA methods. They explain that these methods are based on simple regression, such as yik = ak + bk xi + error,

(Equation 1)

where: yik = abundance of species k in sample i; xi = explanatory variable in sample i; ak = unknown regression coefficient for intercept; bk = unknown regression coefficient for slope. Unconstrained ordination methods such as PCA and CA find values for ak, bk, and xi that best fit a linear model for multiple species (yik) at once. The values for xi are not constrained to measured values. Constrained ordination methods such as RDA and CCA

18

constrain the possible values of xi to linear combinations of the environmental variables, such as xi = c1zi1 + c2zi2 + … + cmzim,

(Equation 2)

where: cm = unknown coefficient for the mth environmental variable; zim = measured value of mth environmental variable for sample i; This value for the explanatory variable xi can be substituted into Equation 1 to obtain Equation 3. yik = ak + bk c1zi1 + bk c2zi2 + … + bk cmzim + error

(Equation 3)

Constrained methods estimate all the unknown values for multiple species simultaneously. This calculates the first axis. Equation 4 is given to calculate multiple axes. yik = ak + bk1 xi1 + bk2 xi2 + … + bks xis + error,

(Equation 4)

where: bks = species score of kth species on sth axis; xis = sample score of ith sample on sth axis; and the samples scores are xis = c1s zi1 + c2s zi2 + … + cjs zij,

(Equation 5)

where cjs = unknown coefficient of jth environmental variable on sth axis.

19

Ordination Diagrams Ordination methods produce ordination diagrams. Ordination diagrams display environmental variables (environmental factors) as arrows, samples (plots or communities) as points, and species as either arrows or points. Some general guidelines that may be useful for interpreting ordination diagrams are provided below. Examples refer to Figure 3.

20

Figure 3. RDA ordination diagram displaying samples (circles), species (triangles), and environmental variables (arrows).

21

1. Distance between sample points is proportional to the dissimilarity between the samples. Sample points that are nearer are more similar than sample points which are farther apart. Example: Sample 12 is more similar to sample 2 than sample 22. 2. The length of environmental variable arrows corresponds to the amount of variation explained by the variable. Example: Soil moisture has a stronger association with community composition than herbaceous ground cover. 3. The correlation between environmental variables, species displayed as arrows, and ordination axes is inversely proportional to the angle between the arrows. Angles < 90° correspond to positive correlation; right angles correspond to no correlation; and angles > 90° correspond to negative correlation. Example: Deciduous midstory is positively correlated with water (angle a) and negatively correlated with soil moisture (angle b). 4. The strength of the relationship between variables displayed as arrows and variables displayed as points can be estimated by drawing a line perpendicular to the arrow through the point. Lines that intersect arrows farther away from plot center indicate a stronger relationship. Example: CRFR is more associated with higher soil moisture than any other species. 5. Arrows display positive values only. Arrows may be extended in the opposite direction to interpret negative relationships (Lepš and Šmilauer 2003). Example: SOCO is negatively associated with soil moisture.

22

METHODS Study Site This research was conducted on the Arkansas Post National Memorial (hereafter, the Post) located in Arkansas County, Arkansas (Figure 4). The Post is owned and operated by the National Park Service and located at the site of a historic French trading post established in 1686 as the first European settlement in Arkansas (Coleman 2002). The Post is a peninsula and consists of 53.8 ha of water and 114.8 ha land. Approximately 19% (22 ha) of the land area of the Post is mowed grass or pavement (Eads 2001). Backwaters of the Arkansas River form the eastern boundary of the Post, and bayous form the western boundary (Figure 5). Soil at the Post is silt loam with slopes ranging from zero to eight percent (United States Department of Agriculture, Natural Resources Conservation Service 2006). Mean and normal climatic conditions for selected months are presented in Table 1 (National Oceanic and Atmospheric Administration Southern Regional Climate Center 2005).

23

Figure 4. Location of the Arkansas Post National Memorial.

24

Figure 5. Vegetation types and sample plot locations at the Arkansas Post National Memorial, 2005-2006.

25

Table 1. Climatic conditions for Dumas, Arkansas in 2006 during sampling periods. March April May June October Annual Mean precipitation (cm) 13.72 12.70 11.51 9.30 10.57 132.44 Mean Temperature (oC) 13.28 17.78 22.22 26.78 17.89 17.50 Normal daily minimum 7.22 11.39 16.39 20.44 11.33 11.67 Temperature (oC) 19.28 24.17 28.28 32.11 24.39 23.28 Normal daily maximum temperature (oC)

26

Vegetation Types Eads (2001) delineated 13 vegetation types on the Post (Figure 5). Each type is briefly described below. Oak/Hickory: Occurred in the southern end of the Post and consisted of water oak (Quercus nigra) and pecan hickory (Carya illinoensis). The midstory was dense and consisted mostly of deciduous holly (Ilex decidua), Japanese privet (Ligustrum japonicum), and other tree saplings that also commonly occurred as trees in the overstory. The understory was well developed and diverse. Characteristic plants include rattan vine (Berchemia scandens), blackberry (Rubus argutus), and several species of grasses. Two sampling plots were placed within this vegetation type. Oak/Pine: Occurred in the interior of the Post and consisted mainly of cherrybark oak (Q. pagoda), post oak (Q. stellata), water oak, and sweetgum (Liquidambar styraciflua). This area was also interspersed with upland associated species such as loblolly pine (Pinus taeda) and Eastern red cedar (Juniperus virginiana). Midstory species typically consisted of trifoliate orange (Poncirus trifoliata), Japanese privet, and numerous other sapling-sized tree species. The understory was well developed and diverse, with herbaceous species such as smartweed (Polygonum hydropiperoides), bedstraw (Galium aparine), and sanicle (Sanicula canadensis). Six sampling plots were placed within this vegetation type. Oak/Mixed species: Occurred in the southern tip of the Post and consisted of typical bottomland hardwood tree species such as pecan hickory, green ash (Fraxinus pennsylvanica), and water oak. The midstory was dense and consisted mostly of deciduous holly, Japanese privet, and other tree saplings that also commonly occured in 27

the overstory. The understory was well developed and diverse. Characteristic plants include rattan vine, blackberry and several species of grasses. One sampling plot was placed within this vegetation type. Burned Oak/Sweetgum: Occurred in the northern portion of the Post and consisted of mixed oak species and sweetgum, with some interspersed Eastern red cedar. It was prescribed burned in 1994. The midstory consisted of Eastern red cedar, deciduous holly, and other sapling-sized tree species also found in the overstory. The understory was well developed, with dense stands of French mulberry (Callicarpa americana). Other understory species include muscidine (Vitus rotundifolia), blackberry, and several species of grasses. Five sampling plots were placed within this vegetation type. Unburned Oak/Mixed species: Occurred mostly in the northern portion of the Post and consisted of mixed oak species and sweetgum with some interspersed Eastern red cedar. The midstory consisted mainly of Eastern red cedar and deciduous holly. The understory was less developed than the Burned Oak/Sweetgum vegetation type, although some areas have well developed stands of French mulberry. Other understory species include muscidine, blackberry, and several species of grasses. Three sampling plots were placed within this vegetation type. Sweetgum: Consisted mostly of sweetgum. The midstory was sparse and consisted of Sweetgum and winged elm (Ulmus alata). The understory was well developed with many species of herbaceous plants. Common species included sanicle, Virginia creeper (Parthenocissus quinquefolia), elephant’s foot (Elephantopus carolinianus), geum (Geum canadense), and several species of grass. One sampling plot was placed within this vegetation type.

28

Sweetgum/Mixed species: Consisted mostly of sweetgum and mixed oak species, with some interspersed Eastern red cedar. The midstory consisted of Eastern red cedar and deciduous holly. The understory was well developed and diverse. Common understory species include muscidine, blackberry, beggar’s lice (Desmodium paniculatum) and several species of grasses. Two sampling plots were placed within this vegetation type. Sweetgum/Oak: Occurred on the north-eastern portion of the Post and consisted of mixed oak species and sweetgum, with some interspersed Eastern red cedar. Tree density separated this vegetation type from other similar vegetation types. The midstory consisted of Eastern red cedar and deciduous holly. The understory was well developed and consisted mainly of rattan vine, blackberry, and French mulberry. Other understory species included muscidine, poison ivy (Toxicodendron radicans), bedstraw, and several species of grasses. Two sampling plots were placed within this vegetation type. Cedar: Occurred in the western interior of the Post and consisted of a relatively small stand of Eastern red cedar and an undeveloped midstory and understory. One sampling plot was placed within this vegetation type. Mowed with trees: Occurred on the western portions of the Post and consisted of primarily post oak, no midstory, and an undeveloped understory. Selected areas of this vegetation type were regularly mowed. No sampling plots were placed within this vegetation type. Mowed without trees: Occurred throughout the Post and consisted of primarily Bermuda grass (Cynodon dactylon), with no overstory, midstory, or understory. This

29

vegetation type was regularly mowed. No sampling plots were placed within this vegetation type. Tall grass: Occurred on the western edge of the Post and overstory consisted mostly of sweetgum, pecan hickory, and black locust (Robina pseudo-acacia). The midstory was sparse, occurred in localized areas only, and consisted mostly of saplingsized sweetgum and black locust. The understory was well developed and diverse. Common herbaceous plants include longleaf sunflower (Helianthus angustifolius), goldenrod (Solidago canadensis), blackberry, and ragweed (Ambrosia artemisifolia). A small portion of this vegetation type was a restored tallgrass prairie consisting of a variety of grass species. Two sampling plots were placed within this vegetation type. Marsh: Occurred primarily along the edge of water in the northwestern portion of the Post, and consisted of vegetation such as cattail (Typha latifolia). No sampling plots were placed within this vegetation type. Plot Establishment To quantify herpetofaunal species presence and abundance and habitat variables that may influence herpetofaunal community structure 23 plots were randomly located within 10 of the 13 vegetation types present on the Post using a geographic information system (Figure 5). Mowed grass and marsh vegetation types were not sampled. Mowed grass was not considered herpetofaunal habitat and the marsh areas were too small to be sampled with the methods used. Two plots (8 and 18 in Figure 5) were not randomly placed to include water to ensure that the water-to-forest habitat was sampled. The number of plots established within each vegetation type was approximately proportional to the area covered by that type (Table 2). 30

Table 2. Vegetation type, area, and number of plots used to sample amphibians, reptiles, and habitat characteristics at the Arkansas Post National Memorial, 2005-2006. Vegetation type Area Number (ha) of plots Oak/Hickory 9.2 2 Oak/Pine 26.1 6 Oak/mixed 2.8 1 Burned Oak/Sweetgum 19.8 5 Unburned Oak/mixed 13.8 3 Sweetgum 1.1 1 Sweetgum/mixed 5.2 2 Sweetgum/Oak 5.9 2 Cedar 2.8 1 Tallgrass 5.9 2 Total 92.6 25

31

Amphibian and Reptile Surveys Area-constrained searches were conducted by trained crews of two or three individuals to sample amphibian and reptile presence and abundance (Petranka et al. 1993, Dupuis et al. 1995, Crosswhite et al. 1999, Adams and Freedman 1999, Hyde and Simons 2001, Shipman et al. 2004, Loehle et al. 2005). Searches were conducted five times (October 2005; March, April, May, and June 2006). Searches were confined to 25m radius circular plots (Figure 6). Each plot was surveyed during daylight hours. Plots were surveyed thoroughly by visually searching vegetation and the ground surface, and by lifting all cover objects (i.e., rocks, logs, and brush). Cover objects were replaced to minimize impact on later surveys. Aquatic habitats were searched visually and with dip nets. Amphibians and reptiles were captured and detained until surveying of that plot was completed to prevent repeated observations. Amphibians and reptiles were detained in one gallon zip top bags or cotton bags. All individuals captured were released at the point of capture. My animal capture and release methods were approved by the University of Arkansas-Monticello Institutional Animal Care and Use Committee (permit number 0304001). Each 25-m radius plot covered an area of 0.196 ha; therefore the 25 plots covered a total area of 4.9 ha. The total area of available vegetation types on the Post was 92.6 ha (Table 2). Therefore, I sampled 5.3% of the terrestrial habitat available to amphibians and reptiles on the Post.

32

Figure 6. Plot configuration used to sample environmental variables and amphibian and reptile presence and abundance at the Arkansas Post National Memorial, 2005-2006

33

Habitat Measurements Data were collected for a variety of environmental factors in order to quantify as much of the variability in the habitats as practical. A total of 59 environmental factors (Table 3), including 35 vegetation features, were measured. Most environmental factors were quantified at three subplots located 25 m from plot center and arranged 120° from each other (Figure 6). Environmental factors measured within the 25-m radius plot included: decay class, diameter at breast height (DBH), height of all snags ≥ 2 m in height and ≥ 10 cm DBH, and percent of area covered by water. Soil moisture was measured at plot center and at the center of all subplots using a soil moisture meter (Field Scout TDR300 Soil Moisture Meter, Spectrum Technologies, Inc.). Vegetation structure was assessed at each subplot by measuring basal area, canopy height, and canopy cover for hardwoods and conifers separately. Basal area was measured using a 2.5 factor metric wedge prism and canopy cover was estimated using a spherical densiometer at each subplot (Higgins et al. 1996). Vegetation density and vertical profiles were quantified for each subplot. Vegetation density was quantified at 0.25 m, 1.25 m, and 2.25 m heights using a 0.25 m2 density board placed in the center of each subplot. Vegetation density was estimated by observing the percentage of density board obscured from 15 m in the direction of plot center (Crosswhite et al. 2004, Perry and Thill 2005). Vegetation profiles were measured at 4 locations 2 m from subplot center and 90° apart at each subplot using a 10 m telescoping pole. The presence of vegetation intersecting the pole at 1 m increments was recorded for hardwoods and conifers.

34

Table 3. Habitat variables sampled within each sampling plot at the Arkansas Post National Memorial, 2005-2006. Location refers to the location in the plot the variable was measured. Code Variable Location Method N/A Temperature (°C) Plot center Thermometer Shrub % cover trees and shrubs < 1m 4-m2 quadrat Ocular estimation Vine % cover vines < 1m 4-m2 quadrat Ocular estimation Dwa % cover down wood 1-10 cm 4-m2 quadrat Ocular estimation diameter Ocular estimation Dw_10up % cover down wood 10-25 cm 4-m2 quadrat diameter (4 decay classes) Dw_25up % cover down wood > 25 cm 4-m2 quadrat Ocular estimation diameter (4 decay classes) Ocular estimation Herb % cover herbaceous vegetation 1-m2 quadrat Rock % cover rock 1-m2 quadrat Ocular estimation 2 Bare % cover bare ground 1-m quadrat Ocular estimation Moss % cover moss 1-m2 quadrat Ocular estimation 2 Litter % cover leaf litter 1-m quadrat Ocular estimation Ldepth Leaf litter depth 1-m2 quadrat Pointed metric ruler DBA_FNL or Basal area (deciduous and Subplot 2.5 BAF metric CBA_FNL conifer) prism DHGT or Canopy height (deciduous and Subplot Clinometer CHGT conifer) DCANC or % canopy cover (deciduous Subplot Spherical CCANC and conifer) densiometer DB Horizontal vegetation density Subplot 0.25- m2 density (3 heights) board D_mid or Vertical vegetation profile Subplot 10m telescoping C_mid (10-1m increments for pole deciduous and conifer) Subplot 10m telescoping pole Soil_H2O Soil moisture Subplot Soil moisture meter N/A Snag type (pine or hardwood) Plot HDW_DIA or Snag DBH (cm) Plot Diameter tape PINE_DIA HDW_HGT or Snag height (m) Plot Clinometer PINE_HGT N/A Snag decay class (4 classes) Plot H2O % cover water Plot Ocular estimation N/A Distance to semi-permanent Plot center 50m tape water (m) N/A Distance to permanent water Plot center 50m tape (m) N/A Distance to ecotone (m) Plot center 50m tape

35

Ground cover was quantified by ocular estimation in two 4-m2 quadrats within each subplot, each containing a nested 1-m2 quadrat (Figure 6). Percent vegetative cover ≤ 1 m tall was estimated for trees and shrubs as well as vines within each 4-m2 quadrat. Percent cover of down wood of 1–10 cm, 10–25 cm, and > 25 cm diameter were also recorded for each 4-m2 quadrat with the two larger groups classified into four decay classes (Figure 7; Maser et al. 1979). Percent cover of herbaceous vegetation, rock, moss, bare ground, and leaf litter were estimated within each 1-m2 quadrat. All ocular estimations were recorded using the Daubenmire cover class scale (Table 4; Daubenmire 1959). Litter depth was measured to the nearest 1 mm using a pointed metric ruler.

36

Figure 7. Log decay classes used to classify down logs at the Arkansas Post National Memorial, 2005-2006. Modified from Maser et al. (1979).

37

Table 4. Daubenmire cover-class scale used to estimate percentage cover at the Arkansas Post National Memorial, 2005-2006. 0-5% 6-25% 26-50% 51-75% 76-95% 96-100% Class 1 Class 2 Class 3 Class 4 Class 5 Class 6

38

Diversity Analysis Species richness indices, Shannon-Weaver diversity indices, and ShannonWeaver evenness indices were used to compare herpetofaunal community structure (Ricklefs 1996, Ricklefs and Miller 1999). All indices were calculated according to their formula for each plot during each sampling period. Indices were then averaged for each plot across all sampling periods by calculating the mean. Species richness index was determined by counting the number of species detected within each sampling plot. The Shannon-Weaver diversity index was calculated as follows:

H = −∑ pi ln pi ,

(Equation 6)

where pi = proportion of individuals in plot of species i. The evenness index associated with the Shannon-Weaver diversity index was calculated as follows: J=

H , ln S

(Equation 7)

where S = species richness index. Gradient Analysis Detrended correspondence analysis (DCA) and detrended canonical correspondence analysis (DCCA) were performed in program CANOCO for Windows, version 4.5 (ter Braak and Šmilauer 2002), to investigate whether the data should be analyzed using linear or unimodal methods. Species data that had been averaged by plot were analyzed using DCA and species and environmental data were analyzed using DCCA. All resulting gradient lengths were < 3.0 standard deviations (Table 5), 39

indicating linear methods should be used for data analysis (ter Braak and Šmilauer 2002, Lepš and Šmilauer 2003). To determine major gradients of separation between plots over time, principal components analysis (PCA) was conducted on species data averaged by plot. Species data were log transformed to reduce the effects of abundant and rare species. Data were centered by sample, which sets a zero mean for each row (Lepš and Šmilauer 2003, Crosswhite et al. 2004). Redundancy analysis (RDA) was conducted on species and environmental data. I examined the correlation matrix and adjusted correlated environmental variables with correlation exceeding r = 0.5 where reasonable. For example, due to high correlation among factors, I calculated the mean from the three heights of horizontal vegetation density data. I calculated the mean of the vertical vegetation profile data between 4 and 10 meters for deciduous and coniferous vegetation independently. Vertical vegetation profile data below 4 meters was discarded due to high correlation with horizontal vegetation density data. Partial RDA was conducted on species and environmental data to identify environmental factors that significantly affected species abundance. Species data were log-transformed to reduce effects of abundant and rare species. Species data were centered by species, which sets the mean value to zero for each species. This transformation is obligatory for RDA and any partial linear ordination method. (Lepš and Šmilauer 2003, Crosswhite et al. 2004). Covariables included: time, date, and month of data collection, air temperature, and distance to water. Environmental factors were selected for inclusion by manual forward selection with significance at α = 0.10.

40

Significance was determined using Monte Carlo permutation tests (499 permutations) with blocks defined by sampling month (Spitzer et al. 1993, ter Braak and Šmilauer 2002, Lepš and Šmilauer 2003, Crosswhite et al. 2004).

41

Table 5. Gradient lengths from detrended analyses expressed as standard deviations. If gradient lengths were <3.0 standard deviations, linear methods were chosen for data analysis (ter Braak and Šmilauer 2002, Lepš and Šmilauer 2003). Gradient length (standard deviation) Analysis Data Axis 1 Axis 2 Axis 3 Axis 4 DCA Species data 1.460 1.616 1.017 0.879 averaged by plot DCCA Species and 2.582 2.460 1.701 1.763 habitat data

42

RESULTS Survey Results I recorded 2934 individuals representing 23 species. Anurans accounted for 95% (n = 2787) and skinks and lizards 2.25% (n = 66) of the total number of individuals recorded. Snakes accounted for 1.87% (n = 55), salamanders 0.58% (n = 17), turtles 0.24% (n = 7), and alligators 0.07% (n = 2) of the total number of individuals recorded. Northern cricket frogs (Acris crepitans) were the most abundant species (Table 6). Species abundance was not evenly distributed across the Post (Figure 8). Species richness and diversity were not evenly distributed on the Post (Table 7, Figure 9, and Figure 10).

43

Table 6. Species codes, common names, scientific names, number recorded, and proportion of total sample. Species Code CRFR GRTR BRFR SLFR ENTO FOTO MASA CENE BRSK GRSK NFLI NGAN RGSN BBWS YEWS DIWS MGWS WRSN RASN WECO SOCO TTBT ANAL

Common Name Northern cricket frogs Green treefrog Bronze frog Southern leopard frog Eastern narrowmouth toad Fowlers toad Marbled salamander Central newt Broadhead skink Ground skink Northern fence lizard Northern green anole Rough green snake Broad-banded watersnake Yellowbelly watersnake Diamondback watersnake Mississippi green watersnake Western ribbon snake Western rat snake Western cottonmouth Southern copperhead Three-toed box turtle American alligator

Acris crepitans Hyla cinerea Rana clamitans Rana spenocephala Gastrophryne carolinensis Bufo fowleri Ambystoma opacum Notophthalmus viridescens Eumeces laticeps Scincella lateralis Sceloporus undulates Anolis carolinensis Opheodrys aestivus Nerodia fasciata Nerodia erythrogaster Nerodia rhombifer Nerodia cyclopion Thamnophis sirtalis Elaphe obsolete Agkistrodon piscivorus Agkistrodon contortrix Terrapene carolina Alligator mississippiensis

44

Individuals Recorded 2700 20 50 5 1 1 1 16 4 57 2 2 6 13 1 3 3 11 3 14 1 7 2

Percent of Total Sample 92.02 0.68 1.70 0.17 0.03 0.03 0.03 0.55 0.14 1.94 0.07 0.07 0.20 0.44 0.03 0.10 0.10 0.37 0.10 0.48 0.03 0.24 0.0007

6

Log(abundance)

5 Crocodilians Turtles Snakes

4 3

Lizard/Skink Salamanders Anurans

2 1 0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Plot

Figure 8. Total abundance (log transformed) of amphibians and reptiles recorded by sampling plots on the Arkansas Post National Memorial, 2005-2006.

45

Table 7. Mean species richness (S), Shannon-Weaver diversity indices (H), and Shannon-Weaver evenness indices (J) for each sampling plot on the Arkansas Post National Memorial, 2005-2006 across all sampling periods. Total Herpetofauna Amphibians Reptiles Plot S H J S H J S H J 1 1.6 0.22 0.19 0.8 0.07 0.10 0.8 0.14 0.20 2 2.0 0.42 0.51 1.4 0.15 0.22 0.6 0.00 0.00 3 1.8 0.27 0.27 0.8 0.00 0.00 1.0 0.25 0.36 4 1.0 0.08 0.11 0.6 0.00 0.00 0.4 0.00 0.00 5 1.2 0.04 0.05 0.8 0.00 0.00 0.4 0.00 0.00 6 0.8 0.00 0.00 0.8 0.00 0.00 0.0 0.00 0.00 7 1.8 0.22 0.27 1.8 0.22 0.27 0.0 0.00 0.00 8 3.8 0.61 0.53 2.4 0.38 0.52 1.4 0.40 0.36 9 1.6 0.37 0.37 0.8 0.00 0.00 0.8 0.22 0.20 10 1.6 0.44 0.55 0.8 0.00 0.00 0.8 0.13 0.18 11 1.0 0.28 0.40 0.2 0.00 0.00 0.8 0.14 0.20 12 1.2 0.11 0.10 1.2 0.11 0.10 0.0 0.00 0.00 13 1.0 0.17 0.16 0.6 0.10 0.14 0.4 0.00 0.00 14 0.6 0.00 0.00 0.4 0.00 0.00 0.2 0.00 0.00 15 2.2 0.44 0.45 1.8 0.38 0.44 0.4 0.14 0.20 16 1.2 0.08 0.12 1.0 0.00 0.00 0.2 0.00 0.00 17 3.8 0.41 0.24 2.2 0.22 0.23 1.6 0.44 0.40 18 4.2 0.26 0.18 2.0 0.10 0.14 2.2 0.54 0.36 19 0.8 0.00 0.00 0.6 0.00 0.00 0.2 0.00 0.00 20 1.0 0.08 0.12 0.8 0.00 0.00 0.2 0.00 0.00 21 1.2 0.18 0.26 0.8 0.00 0.00 0.4 0.00 0.00 22 2.0 0.43 0.54 1.6 0.25 0.36 0.4 0.00 0.00 23 2.0 0.31 0.31 1.6 0.21 0.31 0.4 0.00 0.00 24 3.2 0.99 0.91 1.0 0.13 0.18 2.2 0.55 0.56 25 2.0 0.53 0.55 1.0 0.02 0.03 1.0 0.25 0.36

46

Figure 9. Mean herpetofaunal species richness on the Arkansas Post National Memorial, 2005-2006.

47

Figure 10. Mean herpetofaunal Shannon-Weaver diversity indices plotted by plot for the Arkansas Post National Memorial, 2005-2006.

48

Principal Components Analysis Principle components analysis (PCA) separated my sampling plots along interpretable gradients (Figure 11). The first 3 axes accounted for 97.2% of the variation in species data. Axis 1 accounted for 93.2 % (λ = 0.932), axis 2 accounted for 2.4% (λ = 0.024), and axis 3 accounted for 1.6% (λ = 0.016). Axes 1 and 2 were positively correlated with soil moisture (r = 0.72 and r = 0.56, respectively) and percent cover of water (r = 0.51 and r = 0.20, respectively). Axis 1 was negatively correlated with percent cover of shrubs (r = -0.45) and undergrowth density (r = -0.44). Axis 1 is effectively a moisture gradient extending from xeric plots on the left side of the diagram to more mesic plots on the right (Figure 11). Vegetation cover type was not important in structuring communities. Plot-to-environment relationships were displayed when environmental factors were projected onto the PCA ordination diagram (Figures 12 and 13). PCA does not use environmental factors during analysis; I display them in Figures 12 and 13 to illustrate the underlying role environmental factors may play in structuring herpetological communities at the Post. Environmental factors were selected for inclusion if r > 0.30 for axis 1, 2, or 3. Plots were also separated by deciduous or coniferous vegetation when plotted on axis 3 (Figure 13).

49

Figure 11. PCA ordination diagram of plots sampled at the Arkansas Post National Memorial, 2005-2006.

50

Figure 12. PCA ordination diagram of plots sampled at the Arkansas Post National Memorial, 2005-2006. Axes 1 and 2 are plotted with environmental factors projected.

51

Figure 13. PCA ordination diagram of plots sampled at the Arkansas Post National Memorial, 2005-2006. Axes 1 and 3 are plotted with environmental factors projected.

52

Partial Redundancy Analysis Species-environment relationships were analyzed by partial RDA (Figure 14). The horizontal axis accounted for 24.8 % and the vertical axis accounted for 2.1 % of the species variance. Significance values of environmental factors are located in Table 8. Analysis summary is in Table 9. Amphibians were associated more with soil moisture than reptiles, which were associated more with vegetation structure and standing water. Northern cricket frogs, bronze frogs (Rana clamitans), green treefrogs (Hyla cinerea), and broad-banded watersnakes (Nerodia fasciata) were associated strongly with soil moisture. American alligators (Alligator mississippiensis) were associated with water. Southern copperheads (Agkistrodon contortrix), Western ribbon snakes (Thamnophis proximus), three-toed box turtles (Terrapene carolina), and Mississippi green watersnakes (Nerodia cyclopion) were associated with dense understory and thick deciduous midstory.

53

Figure 14. RDA ordination diagram of species and environmental factors from sampled plots at the Arkansas Post National Memorial, 2005-2006. Environmental factor-to-axes correlations and P-values are in Table 8. Environmental factor codes are in Table 3 Species codes are in Table 6.

54

Table 8. Environmental factor-to-axes correlations and P-values for Figure 14. Axes P-value Variable 1 2 3 4 Soil_ H2O 0.4879 -0.0915 0.1382 0.0842 0.0020 HDW_DIA 0.2455 0.2291 -0.2071 -0.1033 0.0040 Rock -0.0851 -0.11 0.0008 -0.1029 0.0880 DB -0.1569 0.0701 0.0944 0.3603 0.0200 Litter -0.298 0.0639 0.1227 -0.1688 0.0660 H2O 0.3101 0.3942 0.1379 -0.0225 0.0900 D_mid -0.1112 0.2419 -0.1304 0.048 0.0820

55

Table 9. RDA summary for Figure 14. Axes 1 Eigenvalues: 0.162 Species-environment correlations: 0.619 Cumulative percentage variance of species data: 24.8 of species-environment relation: 86 Sum of all eigenvalues: Sum of all canonical eigenvalues:

56

2 0.014 0.529

3 0.007 0.382

4 0.004 0.441

26.9 93.2

27.9 96.9

28.5 98.8

Total 1

0.654 0.189

Reptiles were associated with standing water, but soil moisture was not significant (Figure 15). The horizontal axis accounted for 12.1% and was positively correlated with hardwood and pine snag diameter (Table 10). The vertical axis accounted for 6.2% of the species variation. Percent water coverage was the main environmental factor associated with the vertical axis. Analysis summary is in Table 11.

57

Figure 15. RDA ordination diagram of reptile-environment relationships at the Arkansas Post National Memorial, 2005-2006.

58

Table 10. Environmental factor-to-axes correlations and P-values for Figure 15. Axes P-value Variable 1 2 3 4 DB 0.0258 -0.2406 0.3976 0.0231 0.0700 Dw_10up -0.2547 -0.075 -0.2007 -0.0977 0.0060 Bare -0.0998 -0.1194 -0.0088 0.0049 0.0720 PINE_DIA 0.3386 -0.068 0.0955 -0.2512 0.0200 HDW_DIA 0.3126 0.0413 -0.2211 0.2014 0.0100 H20 0.197 0.4415 0.1488 0.0201 0.0080

59

Table 11. RDA summary for Figure 15. Axes 1 Eigenvalues: 0.095 Species-environment correlations: 0.627 Cumulative percentage variance of species data: 12.1 of species-environment relation: 52.5 Sum of all eigenvalues: Sum of all canonical eigenvalues:

60

2 0.048 0.57

3 0.019 0.551

4 0.008 0.335

18.3 79.1

20.8 89.7

21.8 94

Total 1

0.782 0.181

Amphibians were closely associated with soil moisture (Figure 16). The horizontal axis accounted for 28.2% of species variation and was negatively correlated with soil moisture and positively correlated with leaf litter coverage (Table 12). The vertical axis accounted for 0.6% of species variation and was negatively correlated with hardwood snag diameter. Analysis summary is in Table 13. Northern cricket frogs, bronze frogs, and green treefrogs were positively associated with soil moisture and hardwood snag diameter while negatively associated with thick layers of leaf litter.

61

Figure 16. RDA ordination diagram of amphibian-environment relationships at the Arkansas Post National Memorial, 2005-2006.

62

Table 12. Environmental factor-to-axes correlations and P-values for Figure 16. P-value

Axes Variable Shrub Rock Litter HDW_DIA Soil_H2O

1 0.1936 0.0963 0.3058 -0.2252 -0.4586

2 0.0534 -0.0198 0.0778 -0.1692 0.1274

3 -0.0315 -0.0204 -0.1441 -0.1431 -0.0022

63

4 0.0846 -0.0301 -0.024 0.0104 0.0069

0.0540 0.0640 0.0880 0.0100 0.0020

Table 13. RDA summary for Figure 16. Axes 1 Eigenvalues: 0.169 Species-environment correlations: 0.596 Cumulative percentage variance of species data: 28.2 of species-environment relation: 97 Sum of all eigenvalues: Sum of all canonical eigenvalues:

64

2 0.004 0.259

3 0.001 0.221

4 0 0.095

28.8 99.1

29 99.9

29.1 100

Total 1

0.599 0.174

DISCUSSION Surveys Area constrained searches were used in this study for several reasons: 1) they have been successfully used to collect reptiles and amphibians in many other studies (Petranka et al. 1993, Ash and Bruce 1994, Dodd et al. 1994, Petranka 1994, Dupuis et al. 1995, Block and Morrison 1998, Crosswhite et al. 1999, Adams and Freedman 1999, Hyde and Simons 2001, Shipman et al. 2004, Loehle et al. 2005), 2) they have high repeatability, and 3) they minimize the degradation of habitat (Smith and Petranka 2000). Area constrained searches are favorable over time-constrained searches when habitats to be searched vary in structure (Petranka 1994). Crosswhite et al. (1999) concluded that visual encounter searches, which include area and time constrained methods, are more effective than drift fence arrays for capturing sedentary species that are otherwise difficult to trap. Trapping methods were not used for several reasons. The Post is a relatively small historic site accessible by many foot paths. Park administration requested minimal visual disturbance. There were few locations where large trapping structures such as drift fence arrays could be placed and not seen by park visitors. Moreover, pitfall traps in mesic areas would have filled with water increasing the chance for drowning captured individuals. My area constrained searches did not sample all herpetofaunal species present on the Post, and diurnal searches, of course, fail to detect many nocturnal species, which include many anurans. McCallum et al. (2003) recorded several nocturnal species we failed to detect, including the American toad (Bufo americanus), spring peeper 65

(Pseudacris crucifer), and Cope’s gray treefrog (Hyla chrysoscelis). I also did not extensively sample aquatic species. McCallum et al (2003) recorded six species of aquatic turtles my search methods failed to detect. Species Richness I detected 23 species of amphibians and reptiles on the Post, recording two species that McCallum et al. (2003) failed to detect. These were one Southern copperhead and sixteen central newts (Notophthalmus viridescens louisianensis). The central newt is a new county record for Arkansas County. The addition of these species to the herpetofaunal inventory increases the total recorded species richness for the Post to 39 species. Diversity Indices Areas of high species richness did not correspond to areas with high ShannonWeaver diversity index values. This was due to the high abundance of Northern cricket frogs in these plots. The diversity index was negatively influenced by the unevenness of the abundance data. Communities with high diversity index values contained fewer species; however, their abundances were uniform. This is a common phenomenon when diversity indices are used to analyze biological communities, a phenomenon that limits their usefulness (Ricklefs 1996, Ricklefs and Miller 1999). Communities Communities are defined by species composition in a given area and therefore the individuals within each plot represented a community. Communities were analyzed over time with PCA by averaging all samples for each plot. PCA revealed the importance of 66

the soil moisture to community structure of amphibians and reptiles at this spatial scale (Figure 12). The communities in mesic plots 17 and 18, were comprised of more individuals than the communities in xeric plots, such as 11 and 14 (Figure 8). The importance of moisture was evident by the large variance explained by axis 1 (93.2%). Very little variance was left to be explained by other factors. However, the moisture gradient was not water alone. Correlations between other environmental factors and soil moisture showed that xeric sites had thick understory and thick leaf litter layers. Mesic sites tended to have open understory with large deciduous snags. Partial RDA allowed for examination of species-to-environment associations. Anurans were most closely associated with soil moisture (Figure 14). Broad-banded watersnakes were also associated with soil moisture; however, this may have been in response to presence of anuran prey. Reptiles showed less dependence on soil moisture (Figure 15). Snakes that are traditionally associated with water, such as the Western cottonmouth (Agkistrodon piscivorus), yellowbelly watersnake (Nerodia erythrogaster), diamondback watersnake (N. rhombifer), and broad-banded watersnake had a positive association with water while the rough green snake (Opheodrys aestivus) was less associated with water. Terrestrial species such as three-toed box turtle, Western ribbon snakes, and Southern copperheads had a negative association with water and positive association with thick understory. The importance of moisture structuring amphibian and reptile communities is not surprising. Water and soil moisture are required for large populations of anurans. Frogs require soil moisture to prevent desiccation and standing water for reproduction. Large populations of prey species, such as frogs, near water may account for the higher

67

occurrence of snakes near water. Watersnakes can survive away from standing water and high soil moisture; however, they may be limited by prey availability. Xeric habitats supported species that are less dependent on frogs as their main prey. Western ribbon snakes and Southern copperheads were found in plots that contained no standing water. Rough green snakes are less restricted to mesic habitats because they mainly prey on insect species that are not dependent on high soil moisture and standing water (Mount 1996). Mesic sites near water also had high sunlight penetration. Regions at the Post covered by bayous and backwater do not have many trees. This creates ideal watersnake habitat since there is water allowing escape from predators, frogs and fish to consume, and sunlight available to bask in. Low percentages of variation in species data were explained with RDA. This is common when analyzing abundance data with CANOCO. Other sources of variation could biological interactions and variation associated with sampling methods. However, ter Braak and Šmilauer (2002) claim that ordination diagrams may be valid and informative even if a low percentage of species variance is explained. These results generally agree with other herpetofaunal community studies (Bennett et al. 1980, Pough et al. 1987, Dunson and Travis 1991, Dupuis et al. 1995, Block and Morrison 1998, Hofer et al. 2000, Crosswhite et al. 2004, Loehle et al. 2005), which identified environmental factors that affect soil moisture as significant in structuring amphibian and reptile communities. Crosswhite et al. (2004) concluded that canopy cover, litter depth, woody plant cover, and large down wood, created microhabitat conditions appropriate for amphibians and reptiles. Interestingly, all these environmental

68

features promote soil moisture and high atmospheric humidity. Dupuis et al. (1995) identified downed wood and understory as environmental factors important to salamanders because they promote soil moisture and high humidity. Loehle et al. (2005) found that amphibians were associated with high basal areas because of the high moisture levels in those habitat types. Dense vegetation cover shields the forest floor from direct sunlight promoting high moisture levels. These results indicate that amphibian and reptile communities are structured by environmental variables that affect moisture levels. Some forest management practices, such as clearcutting, drastically change forest structure, and affect the forest’s ability to maintain high moisture levels important to herpetofauna. Although the effects of forest management practices on herpetofauna are not well studied and somewhat controversial, salamanders have been shown to disappear from clearcuts (Petranka 1994, Ash and Bruce 1994). McLeod and Gates (1998) found that amphibians were less abundant in areas where timber harvesting and prescribed burning had occurred. They speculated that these management practices reduced canopy cover and leaf litter depth, which led to higher temperatures and low moisture levels. They concluded that disturbances from forest management could negatively affect amphibian and reptile diversity in local areas. Crosswhite et al. (2004) conducted their study of amphibian and reptile habitat relationships by sampling in areas of three different management treatments. They sampled in clearcuts, natural stands over 80 years old, and selectively harvested areas. They found the herpetofaunal community in each area was significantly different and that clearcuts support fewer amphibians but more reptiles. They concluded that clearcuts

69

attracted reptiles because of increased presence of prey species such as birds, rodents, and insects but that amphibians preferred forests with high moisture levels. Management Recommendations The Post, being a peninsula, supports a diverse community of amphibians and reptiles, partially due to the length of forest-to-water edge habitat. However, large areas of mowed grass are also present at the Post. To maintain and enhance amphibian and reptile communities on the Post, I recommend the following. 1. Protect natural forest-to-water edge habitat. These areas have high species richness. 2. Ephemeral water bodies should be made available in selected locations throughout the Post. Increasing temporary water availability may allow water dependent populations to increase. This could be done by creating small ponds (5-10m diameter) in xeric locations such as areas of burned oak/sweetgum vegetation type in northern portions of the Post. 3. Forest management could protect herpetofaunal habitat by managing for increased down wood, full canopy coverage, and leaf litter coverage. This can be accomplished by continuing to restrict fire and tree cutting. 4. Restrict use of erosion control devices. Plots bordering water but that had shorelines consisting of erosion control devices had lower species abundance than plots with natural shorelines. Consequences of erosion control devices should be carefully considered before decisions are made. 5. Give special management consideration to the Eastern red cedar and sweet gum vegetation types as they were the only areas I detected high population levels of central newts and may be important to this population. 70

LITERATURE CITED Adams, J.D., and B. Freedman. 1999. Comparative catch efficiency of amphibian sampling methods in terrestrial habitats in southern New Brunswick. Canadian Field-Naturalist 113(3):493-496. Ash, A.N., and T.C. Bruce. 1994. Impacts of harvesting on salamanders. Conservation Biology 8(1):300-301. Aubry, K.B. 2000. Amphibians in managed, second-growth Douglas-fir forests. Journal of Wildlife Management 64(4):1041-1052. Barinaga, M. 1990. Where have all the froggies gone? Science 247:1033-1034. Bell, K.E. and M.A. Donnelly. 2006. Influence of forest fragmentation on community structure of frogs and lizards in Northeastern Costa Rica. Conservation Biology 20(6):1750-1760. Bennett, S.H., J.W. Gibbons, and J. Glanville. 1980. Terrestrial activity, abundance and diversity of amphibians in differently managed forest types. American Midland Naturalist 103:412-416. Bestelmeyer, B.T., J.R. Brown, K.M. Havstad, R. Alexander, G. Chavez, and J.E. Herrick. 2003. Development and use of state-and-transition models for rangelands. Journal of Range Management 56(2):114-126. Blaustein, A.R., and D.B. Wake. 1995. The puzzle of declining amphibian populations. Scientific American 272(4):52-58. Blasustein, A.R., D.B. Wake, and W.P. Sousa. 1994. Amphibian declines: judging stability, persistence, and susceptibility and populations to local and global extinctions. Conservation Biology 8(1):60-71. Block, W.M. and M.L. Morrison. 1998. Habitat relationships of amphibians and reptiles in California oak woodlands. Journal of Herpetology 32(1):51-60. Bradshaw, T. and M. Mortimer. 1986. Evolution in communities. Pages 309-341 in J. Kikkawa and D. J. Anderson, ed. Community ecology: pattern and process. Blackwell Scientific Publications, Melbourne. 432 p. Clements, F.E. 1916. Plant succession, an analysis of the development of vegetation. Carnegie Institute of Washington, Washington, D.C. 512 p.

71

Coleman, R.E. 2002. The Arkansas Post story: Arkansas Post National Memorial. Division of History, Southwest Cultural Resources Center, Southwest Region, National Park Service, Department of the Interior. Professional Papers No. 12. Cox, G.W. 1997. Conservation biology: concepts and applications. McGraw Hill Custom Publishing, Boston. 362 p. Cowles, H.C. 1899. The ecological relations of the vegetation of the sand dunes of Lake Michigan. Botanical Gazette. 24:95-117, 167-202, 281-308, 361-91. Crosswhite, D.L., S.F. Fox, and R.E. Thill. 1999. Comparison of methods for monitoring reptiles and amphibians in upland forests of the Ouachita Mountains. Proceedings of the Oklahoma Academy of Science 79:54-50. Crosswhite, D.L., S.F. Fox, and R.E. Thill. 2004. Herpetological habitat relations in the Ouachita Mountains, Arkansas. Pages 273-282 in Guldin J.M. (ed.). Ouachita and Ozark Mounatains symposium: ecosystem management research. U.S. Department of Argiculture, forest Service, Southern Research Station, Asheville, North Carolina. Cushman, S.A. 2006. Effects of habitat loss and fragmentation on amphibians: A review and prospectus. Biological Conservation 128(2):231-240. Daubenmire, R.F. 1959. A canopy-coverage method of vegetational analysis. Northwest Science 33:43-64. DeGraaf, R.M., and D.D. Ruidis. 1990. Herpetofaunal species composition and relative abundance among three New England forest types. Forest Ecology and Management 32:155-165. Dice, L.R. 1968. Natural Communities. Univeristy of Michigan Press, Ann Arbor. 547 p. Dodd, C.K., Jr., R. Franz, and L.L. Smith. 1994. Activity patterns and habitat use of box turtles (Terrapene Carolina bauri) on a Florida island, with recommendations for management. Chelonian Conservation and Biology 1(2):97-106. Dunson, W.A., and J. Travis. 1991. The role of abiotic factors in community organization. The American Naturalist 138(5):1067-1091. Dupuis, L.A., J.N.M. Smith, and F. Bunnell. 1995. Relation of terrestrial-breeding amphibian abundance to tree-stand age. Conservation Biology 9(3):645-653. Eads, G.K. 2001. An assessment of tick density and tick-borne disease frequency at Arkansas Post National Memorial. M.S. Thesis. University of Arkansas at Monticello, Monticello, Arkansas.

72

Gibbons, J.W., D.E. Scott, T.J. Ryan, K.A. Buhlmann, T.D. Tuberville, B.S. Metts, J.L. Greene, T. Mills, Y. Leiden, S. Poppy, and C.T. Winne. 2000. The global decline of reptiles, déjà vu amphibians. BioScience 50(8):653-666. Gleason, H.A. 1926. The individualistic concept of the plant association. Bulletin of the Torrey Botanical Club. 53(1):7-26. Hanlin, H.G., F.D. Martin, L.D. Wike, and S.H. Bennett. 2000. Terrestrial activity, abundance, and species richness of amphibians in managed forests in South Carolina. American Midland Naturalist 143(1):70-83. Herbeck, L.A., and D.R. Larsen. 1999. Plethodontid salamander response to silvicultural practices in Missouri Ozark forests. Conservation Biology 13(3):623-632. Higgins, K.F., J.L. Oldemeyer, K.J. Jenkins, G.K. Clambey, and R.F. Harlow. 1996. Vegetation sampling and measurement. Pages 567-591 in Bookhout, T.A. (ed.) Research and Management Techniques for Wildlife and Habitats. The Wildlife Society, Bethesda, Maryland. 740 p. Hofer, U., L.F. Bersier, and D. Borcard. 2000. Ecotones and gradient as determinants of herpetofaunal community structure in the primary forest of Mount Kupe, Cameroon. Journal of tropical ecology 16(4):517-533. Hyde, E.J., and T.R. Simons. 2001. Sampling plethodontid salamanders: sources of variability. Journal of Wildlife Management 65(4):624-632. Knutson, M.G., J.R. Sauer, D.A. Olsen, M.J. Mossman, L.M. Hemesath, M.J. Lannoo. 1999. Conservation Biology 13(6):1437-1446. Laycock, W.A. 1991. Stable states and thresholds of range condition on North American rangelands: a viewpoint. Journal of Range Management 44(5):427-433. Lepš, J and P Šmilauer. 2003. Multivariate analysis of ecological data using CANOCO. Cambridge University Press, Cambridge. 269 p. Loehle, C., T.B. Wigley, P.A. Shipman, S.F. Fox, S. Rutzmoser, R.E. Thill, M.A. Melchiors. 2005. Herpetofaunal species richness response to forest landscape structure in Arkansas. Forest Ecology and Management 209:293-308. Losos, J.B., K.I. Warheit, and T.W. Schoener. 1997. Adaptive differentiation following experimental island colonization in Anolis lizards. Nature 387:70-73. Marsh, D.M. and P.C. Trenham. 2001. Metapopulation dynamics and amphibian conservation. Conservation Biology 15(1):40-49.

73

Maser, C., R.G. Anderson, K. Cormack, Jr., J.T. Williams, and R.E. Martin. 1979. Dead and down woody material. Pages 78-95 in J.W. Thomas, ed. Wildlife habitats in managed forests: the Blue Mountains of Oregon and Washington. U.S. Department of Agriculture Forest Service, Agriculture handbook No. 553. Matlack, G. 1994. Plant demography, land-use history, and the commercial use of forests. 1994. Conservation Biology 8(1):298-299. McCallum, L.M., S.E. Trauth, R.G. Neal, and V. Hoffman. 2003. A herpetofaunal inventory of Arkansas post National Memorial, Arkansas County, Arkansas. Journal of the Arkansas Academy of Science 57:122-130. McLeod, R.F. and J.E. Gates. 1998. Response of herpetofaunal communities to forest cutting and burning at Chesapeake Farms, Maryland. American Midland Naturalist 139(1):164-177. Mount, R.H. 1996. The reptiles and amphibians of Alabama. University of Alabama Press, Tuscaloosa. 347 p. National Oceanic and Atmospheric Association, Southern Regional Climate Center. 2004. SRCC climate atlas. Accessed 2005 December 01. National Parks Omnibus Management Act of 1998 Odum, E.P. 1971. Fundamentals of ecology (third ed.). W. B. Saunders company, Philadelphia. 574 p. Palmer, M.W. 1993. Putting things in even better order: the advantages of canonical correspondence analysis. Ecology 74:2215-2230. Pechmann, J.H.K., D.E. Scott, R.D. Semlitsch, J.P. Caldwell, L.J. Vitt, and J.W. Gibbons. 1991. Declining amphibian populations: the problem of separating human impacts from natural fluctuations. Science 253:892-895. Perry, R.W. and R.E. Thill. 2005. Small-mammal responses to pine regeneration treatments in the Ouachita Mountains of Arkansas and Oklahoma, USA. Forest Ecology and Management 219:81-94. Petranka, J.W. 1994. Response to impact of timber harvesting on salamanders. Conservation Biology 8(1):302-304. Petranka, J.W., M.E. Eldridge, and K.E. Haley. 1993. Effects of timber harvesting on southern Appalachian salamanders. Conservation Biology 7(2):363-370.

74

Pough F.H., E.M. Smith, D.H. Rhodes, and A. Collazo. 1987. The abundance of salamanders in forest stands with different histories of disturbance. Forest Ecology and Management 20:1-9. Pounds, J.A., and M.L. Crump. 1994. Amphibian declines and climate disturbance: the case of the golden toad and the harlequin frog. Conservation Biology 8(1):72-85. Ricklefs, R.E. 1996. The economy of nature (fourth edition). W.H. Freeman and Company, New York. 678p. Ricklefs, R.E., and G.L. Miller. 1999. Ecology (fourth edition). W.H. Freeman and Company, New York. 896p. Semlitsch, R.D., T.J. Ryan, K. Hamed, M. Chatfield, B. Dreshman, N. Pekarek, M. Spath, and A. Watland. 2007. Salamander abundance along road edges and within abandoned logging roads in Appalachian forests. Conservation Biology 21(1):159-167. Shelford, V.E. 1931. Some concepts of bioecology. Ecology 12:445-467. Shipman, P.A., S.F. Fox, R.E. Thrill, J.P. Phelps, and D.M. Leslie, Jr. 2004. Reptile Communities under diverse forest management in the Ouachita Mountains, Arkansas. Pages 174-182 in Guldin J.M. (ed.). Ouachita and Ozark Mounatains symposium: ecosystem management research. U.S. Department of Argiculture, forest Service, Southern Research Station, Asheville, North Carolina. Skelly, D.K., E.E. Werner, and S.A. Cortwright. 1999. Long-term distributional dynamics of a Michigan amphibian assemblage. Ecology 80(7):2326-2337. Smith, C.K., and J.W. Petranka. 2000. Monitoring terrestrial salamanders: repeatability and validity of area-constrained cover object searches. Journal of Herpetology 34(4):547-557. Spitzer, K., V. Novotny, M. Tonner, and J. Leps. 1993. Habitat preferences, distribution and seasonality of the butterflies in montane tropical rain forest, Vietnam. Journal of Biogeography 20:109-121. Tansley, A.G. 1935. The use and abuse of vegetational concepts and terms. Ecology 16(3):284-307. ter Braak, C.J.F. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67:1167-1179. ter Braak, C.J.F., and P. Šmilauer. 2002. CANOCO reference manual and User’s guide to Canoco for Windows: Software for Canonical Community Ordination (version 4.5). Microcomputer Power, Ithaca, New York. 75

Trauth, S.E., H.W. Robinson, and M.V. Plummer. 2004. The ampibians and reptiles of Arkansas. University of Arkansas Press, Fayetteville. 421 p. United States Department of Agriculture, Natural Resources Conservation Service. 2006. Soil survey of Arkansas County, Arkansas. Web Soil Survey 1.1: National Cooperative Soil Survey. Accessed 2006 April 26. Vitt, L.J., E.R. Pianka, W.E. Cooper, Jr., and K. Schwenk. 2003. History and the global ecology of squamate reptiles. The American Naturalist 162(1):44-60. Von Liebig, J. 1840. Die organische chemie in ihrer anwendung auf agrikultur und physiologie. Friedrich Vieweg, Braunschweig, Germany. Wake, D.B. 1991. Declining amphibian populations. Science 253:860. Watt, A.S. 1964. The community and the individual. The Journal of Animal Ecology. 33:203-211. Whiles, M.R., K.R. Lips, C.M. Pringle, S.S. Kilham, R.J. Bixby, R. Brenes, S. Connelly, J.C. Colon-Gaud, M. hunte-Brown, A.D. Huryn, C. Montgomery, and S. Peterson. 2006. The effects of amphibian population declines on the structure and function of Neotropical stream ecosystems. Frontiers in Ecology and the Enviroment 4(1):27-34. Whittaker, R.H. 1953. A consideration of climax theory: the climax as a population and pattern. Ecological Monographs 23(1):41-78. Whittaker, R.H. 1967. Gradient analysis of vegetation. Biological Reviews of the Cambridge Philosophical Society. 42:207-264. Zug, G.R., L.J. Vitt, and J.P. Caldwell. 2001. Herpetology: an introductory biology of amphibians and reptiles. Second Edition. Academic Press, San Diego and London. 630 p.

76