Citation
Movements and Habitat Use of the Santa Rosa Beach Mouse (Peromyscus Polionotus Leucocephalus) in a Successional Dune Mosaic

Material Information

Title:
Movements and Habitat Use of the Santa Rosa Beach Mouse (Peromyscus Polionotus Leucocephalus) in a Successional Dune Mosaic
Creator:
SMITH, KATHRYN ENGA LOUISE
Copyright Date:
2008

Subjects

Subjects / Keywords:
Beaches ( jstor )
Connectivity ( jstor )
Dunes ( jstor )
Habitat conservation ( jstor )
Habitat fragmentation ( jstor )
Landscapes ( jstor )
Mathematical variables ( jstor )
Mice ( jstor )
Species ( jstor )
Vegetation ( jstor )
Santa Rosa Beach ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Kathryn Enga Louise Smith. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
5/1/2005
Resource Identifier:
84053730 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

MOVEMENTS AND HABITAT USE OF THE SANTA ROSA BEACH MOUSE ( Peromyscus polionotus leucocephalus ) IN A SUCCESSIONAL DUNE MOSAIC By KATHRYN ENGA LOUISE SMITH A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003

PAGE 2

Copyright 2003 by Kathryn Enga Louise Smith

PAGE 3

ACKNOWLEDGMENTS Funding for this research was provided through the West Florida Research and Education Center, Institute of Food and Agricultural Sciences, University of Florida. I would like to thank my committee chairs, Drs. Deborah Miller and Mack Thetford, for their encouragement and guidance; and my committee members, Drs. Lyn Branch and George Tanner, for their time and advice with the project. Additional thanks are extended to the staff and personnel of the West Florida Research and Education Center and the Department of Wildlife Ecology and Conservation. I also wish to thank the U.S. Fish and Wildlife Service for providing traps and other supplies, in particular, Lorna Patrick and Bill Lynn. I greatly acknowledge the professional and personal support of Randy Swilling and Brittany Bird, for their continual advice throughout the project. Field assistance was generously provided by Mica Schneider, Julianne Granton, Jennifer DuPree, Jason Liddle, and Jason Tritt. Statistical advice was provided by Andrea Booth, Youngsung Joo, and Dr. Kenneth Portier, who were all generous with time and patience. I owe much appreciation to Jackson Guard, Eglin Air Force Base, for providing access to base property, logistical assistance, and housing throughout the project; in particular, Bob Miller and Bruce Hagedorn provided considerable support. I also wish to thank Eglin Air Force Base firefighters at Station Seven for their generosity and kindness. Finally, I am extremely grateful for the support of my family, Jim, Darlene, James, and Otto, and my very dear friend, Kurt, to whom this work is dedicated. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iii LIST OF TABLES...............................................................................................................v LIST OF FIGURES...........................................................................................................vi ABSTRACT......................................................................................................................vii INTRODUCTION...............................................................................................................1 METHODS..........................................................................................................................5 Study Area....................................................................................................................5 Study Species................................................................................................................6 Trapping Procedure......................................................................................................8 Landscape and Vegetation Protocol.............................................................................9 Statistical Analyses.....................................................................................................12 Movement............................................................................................................12 Habitat Use..........................................................................................................13 Mean Gap or Interdunal Distance........................................................................15 RESULTS..........................................................................................................................17 DISCUSSION....................................................................................................................24 APPENDIX RESEARCH DATA.....................................................................................29 LIST OF REFERENCES...................................................................................................34 BIOGRAPHICAL SKETCH.............................................................................................38 iv

PAGE 5

LIST OF TABLES Table page 2-1 Eleven variables illustrating the variation of dune fragment isolation and connectivity among six study sites...........................................................................15 3-1 Pearson correlation coefficients among habitat patch variables..............................19 3-2 Results of the multivariate Poisson regression procedure testing for the effect of habitat structure on number of first captures for individual trap locations..............20 3-3 The control variable (I se ) for habitat use model, varying by season of trapping session and landscape element.................................................................................20 3-4 Due to a greater than expected use of swale habitat, connectivity value was recalculated for the sites within the Winter and Spring trapping sessions to include swale landscape element as “joined” with other landscape elements.........23 A-1 List of food plants used in calculating percent cover of food species......................29 A-2 Pearson correlation coefficients for variables used in landscape analysis...............30 A-3 Pearson correlation coefficients for microhabitat variables used in habitat selection analyses.....................................................................................................31 A-4 Number of individual mice captured per site, by trapping session and gender (M = Male, F = Female)...........................................................................................33 v

PAGE 6

LIST OF FIGURES Figure page 2-1 Map of Santa Rosa Island, Eglin Air Force Base, Florida with location of six study sites.................................................................................................................14 2-2 Black and white satellite imagery (IKONOS, May 2001) depicting an aerial view of Site E showing placement of the trapping grid among five landscape elements....................................................................................................................14 2-3 Sample 5x5 grid demonstrates the calculation of the connectivity value for study sites...........................................................................................................................16 3-1 Linear regression of mean square movement (MSD) and habitat patch variables: amount of regenerating dune habitat........................................................................21 3-2 Mean movement (SE) for six levels of dune fragment isolation and fragmentation...........................................................................................................22 3-3 Mean number of first captures (SE) per trap..........................................................22 3-4 Linear regression of mean square movement (MSD) and the new connectivity value that incorporates swale as “joined” habitat.....................................................23 A-1 Linear regression of mean square movement (MSD) and habitat patch variables: amount of regenerating dune habitat........................................................................32 vi

PAGE 7

vii Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science MOVEMENTS AND HABITAT USE OF THE SANTA ROSA BEACH MOUSE ( Peromyscus polionotus leucocephalus ) IN A SUCCESSIONAL DUNE MOSAIC By Kathryn Enga Louise Smith May 2003 Chair: Deborah L. Miller Cochair: Mack Thetford Major Department: Wildlife Ecology and Conservation I studied the effects of habitat isolati on and connectivity on movement and habitat use of the Santa Rosa beach mouse ( Peromyscus polionot us leucocephalus ) in a successional dune mosaic. Capture-recapture da ta were examined for six trapping grids selected to represent a range of dune habita t isolation and connectivity. Analyses were focused on determining the influence of connectivity on movement and the influence of micro-habitat elements on habitat use. Mean squared distance is a variance estimate of the locations where individuals were captu red and was used to quantify movement. Habitat use was quantified as the number of first captures for any given trap. Vegetation and landscape data on trap locations were used to characterize the trapping grid and examine preferences within the grid. Move ment distance increased in less connected, more isolated dune patches. Habitat us e varied between seasons and differed by landscape element type, although beach mice dem onstrated a preference for habitat with

PAGE 8

greater vegetation cover and decreased gap distance. Examining both results suggests that increased movement in more isolated habitat may be due to an increase in distances between resources and not due to predilection for gap and interdunal habitat. Since beach mice may be required to move further in fragmented habitat and the overall amount of preferred habitat for beach mice is reduced, habitat fragmentation and isolation may affect beach mice survival and longevity. Removal of these microhabitat variables, as is characteristic of isolated habitats, may have negative consequences for beach mice. viii

PAGE 9

INTRODUCTION Knowledge of how animals move in relation to landscape structure is important when designing landscape-level conservation strategies. Animal movement affects many aspects of a species’ biology, as well as the organization and regulation of populations (Stamps et al. 1987, Wiens et al. 1993, Ims 1995). Research efforts in landscape ecology have demonstrated that structure can influence movement, both on the micro(Kaufman et al. 1983, Barnum et al. 1992, Planz and Kirkland 1992, McMillan and Kaufman 1995) and macro-habitat scale (Kaufman et al. 1983, Wegner and Merriam 1990). Landscape elements can obstruct or facilitate movement, and therefore, have serious implications for population conservation (Kozakiewiez 1993). Improved conservation planning requires knowledge of landscape pattern and the response of organisms to changes in pattern. This is particularly true for species under the stresses of habitat loss and fragmentation (Fahrig and Merriam 1994). Habitat fragmentation results in a landscape where habitat patches are isolated by non-habitat or a restricted habitat matrix (Saunders et al. 1991). The degree of habitat isolation is determined by the species response to landscape structure and the spatial arrangement of landscape elements. Landscape connectivity is a measure of habitat isolation and focuses on how landscape elements influence the ability of organisms to move within the landscape and between habitat patches (Pither and Taylor 1998). Connectivity refers to the arrangement of landscape elements and linkage among habitat patches (Merriam 1984, Bowne et al. 1999) suggesting a “degree to which the landscape 1

PAGE 10

2 facilitates or impedes movement among resource patches” (Taylor 1993). Landscape connectivity is an important characteristic of landscape-organism interactions (Fahrig and Merriam 1985). Understanding the influence of connectivity begins with a focus on how landscape elements influence movement and behavior of organisms. Previous connectivity research suggests critical thresholds exist for species persistence in fragmented landscapes (With and Crist 1995). As connectivity becomes more reduced, species ability to move through and between habitats is reduced in a nonlinear fashion. The critical threshold results when the response produces an abrupt or dramatic change in the relationship. In the case of landscape connectivity and movement, the habitat becomes less interconnected and movement is severely reduced. Understanding the influence of connectivity on animal movement and identifying a connectivity threshold can provide essential data to plan for habitat restoration and conservation. Investigations in habitat fragmentation have primarily been in human altered (Szacki and Liro 1991, Wauters et al. 1994, Delin and Andrn 1999) or manipulated landscapes (Diffendorfer et al. 1995, Collinge 2000), but fragmentation can also be caused by natural disturbances (Kozakiewiez 1993). Natural disturbances create a complex landscape where remnant habitat patches are surrounded by a mosaic of regenerating habitat and non-habitat. The result is a successional mosaic with a patchy distribution of habitat types and varying degrees of isolation (Kincaid and Cameron 1985, Foster and Gaines 1991). The ability of an organism to use a patch or landscape element will be determined by the biophysical nature of the route and the behavior and biology of

PAGE 11

3 the organism (Taylor 1993). Therefore, patch isolation may be examined through a focus on the structure of regenerating habitat and the movement of an organism. Two primary types of movement have been identified: within home-range activity and long-range dispersal. Fragmentation and habitat isolation can affect organism dispersal throughout landscapes. Fragmentation may also affect daily movements, particularly if patches are smaller than a species’ home range or do not contain all required resources. Many studies have focused on microhabitat structural influences on home-range activity (Kaufman et al. 1983, Barnum et al. 1992, Planz and Kirkland 1992, McMillan and Kaufman 1995) and on the influence of fragmentation on dispersal (Wauters et al. 1994, Brooker et al. 1999), but few studies have sought to understand the influence of habitat isolation and connectivity on movement at the scale of the habitat patch. The focus of this study was to examine the influence of fragmentation and patch isolation on movement within a successional habitat mosaic. In addition, we examined the influence of structure at two scales, movements at the habitat patch and the microhabitat scale. Trapping field experiments were conducted to test the influence of landscape structure on the movement behavior of the Santa Rosa beach mouse (Peromyscus polionotus leucocephalus). The Santa Rosa beach mouse resides in a successional mosaic of hurricane-fragmented dune remnants and regenerating dune patches. The fragmented dune landscape provides an ideal setting for studying the response of fauna to landscape structure and fragmentation. The facilitation of regenerating dunes to movement between dune fragments or among habitats is poorly understood, but could provide critical information for conservation and restoration planning. Preliminary observations indicate

PAGE 12

4 that beach mice move between dune fragments and regenerating dune patches, and that pattern of movement is strongly influenced by distance and configuration of regenerating dunes. Assuming that dune structure and vegetation affect movement across a landscape, beach mouse movement should be correlated with sand gap distances and vegetation cover. The goals of this study were to focus on patch isolation and its influence on movement, identify a threshold where organism movement is reduced due to a loss of species perceived connectivity, and suggest which structural characteristics may be crucial to movement. I hypothesized that habitat structure and fragment isolation will influence beach mouse movement and ultimately their ability to move throughout the dune landscape. My objectives were to (1) compare beach mouse movement in several landscapes with contrasting levels of dune fragment isolation and connectivity, (2) determine the influence of dune fragment structure on movement, and (3) determine the influence of microhabitat structure on habitat use and selection.

PAGE 13

METHODS Study Area The study was located in the natural dune landscape of Santa Rosa Island, an approximately 46-km long and -km wide barrier island, located in northwest Florida on the southern border of Escambia, Santa Rosa, and Okaloosa counties. The Gulf of Mexico to the south and Santa Rosa Sound to the north border the island. Study sites were located on Eglin Air Force Base, east of Navarre Beach (3018’ N, 8716’ W; Figure 2-1). This portion of Eglin Air Force Base property is approximately 13-km in length and includes the full width of the island. Development is restricted to a two-lane paved road and several military installations. Santa Rosa Island is a narrow barrier island with a strong linear arrangement, parallel to the coastline, which can be divided into three zones: beach and frontal zone, interior zone, and bay zone. Within these zones, five distinct habitat types have been identified (Chafin et al. 1997). Three of the habitats are found on dune ridges: foredunes, scrub dunes, and coastal grasslands. Foredunes, found within the frontal zone, are the first row of dunes parallel to the Gulf. Foredune vegetation is predominately sea oats (Uniola paniculata L.), with cakile (Cakile spp.) and seashore elder (Iva imbricata Walt.) on the beach-side slope. Scrub dunes can occur in either the frontal or interior zone. When foredunes have been lost to storm impacts, remnant scrub dunes may be the first set of dunes parallel to the beach. Scrub dunes range from 2-10m in height and are characterized by a diverse woody complex (Chrysoma pauciflosculosa (Michx.) Greene, 5

PAGE 14

6 Ceratiola ericoides Michx., Quercus virginiana Mill., Smilax spp.) mixed with dune grasses. Grasses and forbs, such as Maritime bluestem (Schizachyrium maritimum (Chapman) Nash), and Sea beachgrass (Panicum amarum Ell.), dominate flat to gently undulating coastal grasslands found within the interior zone (henceforth, called interior dunes). Interdunal swales are often ephemeral wetlands located as pockets within the interior zone (Spartina patens (Ait.) Muhl, Xyris spp., Juncus spp., Polygala spp.). Mesic flatwoods are a forested habitat type (Pinus elliottii Englem., Quercus virginiana, Ilex spp., Myrica cerifera L.) and are common within the bay zone. Santa Rosa Island has recently been impacted by several hurricanes, of which Hurricane Opal in 1995 was the most damaging. The frontal zone of the island was severely impacted by hurricane and tropical storm activity. Storm surges removed nearly all foredunes, and fragmented and isolated remaining dunes (Figure 2-2). Interior dunes were also damaged leaving frontal dune fragments isolated from each other and interior dune habitats by bare sand gaps. Currently, natural dune regeneration is occurring within these gaps, however slowly. The regenerating dunes may connect fragments to more stable interior scrub and coastal grassland habitats, providing landscape continuity for dune fauna. Study Species The Santa Rosa beach mouse (Peromyscus polionotus leucocephalus) is one of eight subspecies of the Old field mouse (P. polionotus) that reside in the coastal dunes of Florida and Alabama. These populations are small and isolated, making them susceptible to extinction. The added stresses of coastal development, introduced species, and a naturally dynamic environment have aggravated this vulnerability and are cited as primary causes of dwindling populations (Gore and Schaefer 1993a). Currently, the

PAGE 15

7 Endangered Species Act protects seven of the eight subspecies, with one of these apparently extinct. The Santa Rosa beach mouse is the only beach mouse subspecies not listed as endangered, threatened, or as species of special concern by the U.S. Fish and Wildlife Service or the Florida Fish and Wildlife Conservation Commission (FWC). The large tracts of relatively undeveloped property on Santa Rosa Island (predominately located within Gulf Islands National Seashore and Eglin Air Force Base) are cited as the primary reason for the unlisted status of the Santa Rosa beach mouse (Gore and Schaefer 1993b). Although these large tracts are currently protected against commercial or residential development, there is limited knowledge of population recovery following severe storms and whether abrupt habitat loss due to hurricane disturbance affect populations. An understanding of how the fragmented dune structure influences beach mouse populations is needed. Optimal habitat for Gulf Coast subspecies of beach mice is currently designated as foredunes (Holler 1992a). Recent studies have demonstrated that scrub dunes and coastal grasslands are also important habitat types and may be important refugia from storms or supplement habitat when foredunes are damaged (Swilling et al. 1998, Lynn 2000, Sneckenberger 2001). Therefore, movement between and within these habitats may be crucial to species survival, particularly when foredune habitat is damaged following storms. This study focused on the frontal zone dune fragments and associated regenerating dune habitat interior to dune fragments. During this study, dune fragments were treated as habitat patches because their characteristics match those required for the support of burrow sites (Lynn 2000). Regenerating dunes are considered resource patches because they may be limited to foraging or to facilitate movement, but could also host limited

PAGE 16

8 burrow sites. The bare sand between dunes (bare sand or interdunal gaps) is uninhabitable and may to some degree act as a barrier to movement. Swales are frequently wet depressions and are not considered habitat in species descriptions (Holler 1992a, b). Trapping Procedure Field-work was conducted from March to October 2001. Six sites were selected for placement of trapping grids. Each site was associated with a frontal dune fragment through a zone of regenerating or remnant interior dunes into the more continuous coastal grassland habitat. The amount of dune habitat and interdunal gap distances were identified as measurable characteristics affecting dune fragment isolation and connectivity to interior dunes. Therefore, sites were selected with differing amount and configuration of regenerating and remnant interior dunes, gap distances, and vegetation cover. The goal was to select six examples within a range of landscape connectivity. A trapping grid was established at each site using a meter tape and compass. Grids consisted of 91 traps (7 x 13) set at 15-m spacing. Approximately two hectares were trapped per grid. Trapping grids were situated with the southwest corner trap placed randomly within the frontal dune zone on a dune fragment. The remaining traps were measured from this first trap (Figure 2-2). One small Shermin live trap was placed at each trap location. Universal Transverse Mercator (UTM) coordinates for each trap location were collected using a Global Positioning System (GPS) Trimble Asset Surveyor (Trimble Navigation, Ltd.). Traps were set in late evening, near the time of sun down, and baited with oats. Evenings when temperatures were expected to drop below 19C, cotton batting was placed in traps for insulation. Cotton batting was also used when weather conditions predicted evening dew or fog. Traps were not baited if fire ants were present or were checked early in the

PAGE 17

9 evening to prevent predation by ants. Traps were checked and mice released prior to sunrise. All beach mice captured were individually marked by toe-clipping. Individuals were sexed, weighed (g), and placed in an appropriate age class (subadult or adult). Females were identified as pregnant, pregnant and lactating, lactating, or non-lactating. Males were identified as having abdominal or distended testes. Trapping grid and trap location identified capture location. Trapping grids were trapped for four to five consecutive nights, whenever possible. When rain, extreme temperatures, or military operations prevented trapping, the session was resumed when conditions permitted. When possible, all sites were trapped during the same moon phase, avoiding full moon nights where moon fractions were three-quarters to full. Trapping session 1 was conducted during March (henceforth, referred to as the Winter trapping session) and included sites A, B, C, and D. Trapping session 2 was conducted during April May (Spring) and included sites A, B, C, D, and E. Trapping session 3 was conducted during June July (Summer) and included all six sites. Trapping session 4 included all six sites and was conducted during September October (Fall). Landscape and Vegetation Protocol Landscape and vegetation data were collected for each site and associated with trap locations (n = 546). Each trap was distinguished as falling in one of four distinct landscape elements. Landscape elements identified include remnant dune fragment, regenerating dune or remnant interior dune, swale, and interdunal sand gap. The location of dune fragments were delineated and identified following Hurricane Opal using GPS. Regenerating and remnant interior dunes were identified via vegetation and elevation change (sand deposition and aggregation associated with dune grass vegetation). Swales

PAGE 18

10 were depressions and were seasonally wet or wet following storms. Interdunal gaps were devoid of vegetation and had minimal or no sand accumulation. Each trap was found either within vegetation cover or no cover. For all traps within no vegetative cover, distance to a vegetation patch was determined by measuring the distance to the closest vegetation using a meter tap. Traps located within vegetation were given a distance value of zero. Plant species composition and vegetation cover were determined via vegetation transects. Thirteen 105-m transects were assessed for each grid. Each 105-m transect was divided into seven 15-m segments. Each segment was associated with a trap location. For each transect segment, total cover distance for each species was measured. In addition, all bare sand distances greater than 10 cm and interdunal gaps were measured. These gap distances were used to calculate total percent cover of vegetation by subtracting the sum of the gap distances from the total length (1500 cm) and dividing by the total length. Landscape and vegetation data were used to calculate descriptive variables for both individual traps (microhabitat) and the trapping grid as a whole (habitat). Habitat variables were calculated using data for all traps within a trapping grid (Table 2-1). Variables were chosen to highlight structural characteristics of the dune fragment and to describe the level of dune fragment isolation. Dune fragment size and perimeter were calculated from delineated dune fragments. Species richness and vegetation cover variables were calculated from vegetation transects. Landscape composition variables (proportion of swale, regenerating or remnant interior dune landscape elements) were calculated as the proportion of traps located within that element. The eleventh descriptive variable, connectivity value, was designed to describe the level of landscape connectivity for each site. Connectivity value was measured similar to

PAGE 19

11 connectance index, a landscape metric designed to quantify landscape structural connectivity for analysis of spatial data (McGarigal 1995). The connectance index was designed to identify the amount of functional joinings between cells of a landscape. A modified connectivity value developed for this study allowed for a similar quantification of the amount of “functional joinings” between traps (cells). Traps were designated as being within habitat (dune fragment, regenerating or remnant interior dune) or non-habitat (swale and interdunal gap). If a particular trap was located within habitat, each adjacent trap was identified as either being joined (also within habitat) or unjoined (within non-habitat). Connectivity value was calculated as the proportion of joined traps to possible joinings (Figure 2-3). All traps located within non-habitat were given a value of zero, as there were no functional joinings between traps. Microhabitat variables were calculated for each trap location using associated vegetation transects and landscape data. The following microhabitat variables were created: landscape element, distance to a vegetation patch, species richness, total percent cover of vegetation, total percent cover of food species, total percent cover of grass species, total percent cover of non-grass species, largest gap or interdunal distance, and average gap or interdunal distance. Movement Index Trapping data were used to calculate the mean squared distance (MSD) (Slade and Swihart 1983, Slade and Russel 1998) as an estimate of animal movement for each trapping grid and trapping session, mnyyxxMSDniimjniii11122 ,

PAGE 20

12 where x i and y i are individual trap location coordinates, x and y are the mean of the x and y coordinates, n is the number of captures for an individual, and m is the total number of individuals. Individuals that were captured at least twice during the trapping session were used in calculating MSD. Statistical Analyses All statistical analyses were performed using SAS/STAT software Version 8.0 (SAS Institute 1999). Data are reported as mean ( 1 SE). P-values < 0.05 were considered significant on all tests. Movement A multiple regression was used to test the influence of dune fragment structure, isolation, and landscape connectivity on beach mouse movement (MSD). Correlations among habitat variables were calculated using Pearson correlation coefficient. The resulting correlations and graphs of the data were used to select four habitat variables (out of the eleven original habitat variables described in Table 2-1) to test for significance on MSD. The four variables included in analyses were dune fragment area, dune fragment perimeter, connectivity value, and amount of interior dune development. Trapping session was also included in the analysis, for a total of five variables. To find the best-fitting regression equation, a backward elimination procedure was used to reduce the model. The procedure removed the trapping session variable, so the SAS procedure, PROC MIXED, was used to verify the effect of trapping session on MSD (SAS Institute 1999). The result confirmed trapping session was not significant (P = 0.58) and was ignored in results. A connectivity threshold will be determined by a curved trend, or a change in movement based on the level of connectivity.

PAGE 21

13 Habitat Use Habitat use was measured by counting the number of first captures (first unique trap-mouse combination) (HAB) at each trap location. Correlations between microhabitat variables were calculated using Pearson correlation coefficient (Appendix A). When a pair of variables was highly correlated (r 0.70), only one variable was retained for analysis. Habitat use was regressed against the selected microhabitat variables using a multiple Poisson regression. The selected variables included were: landscape element, distance from vegetation patch (m), species richness, percent total cover of vegetation, cover of food species, cover of grasses, cover of herbaceous or woody species, and average gap distance. Trapping session and site were also included as variables within the analysis. The full model was reduced by removing variables individually (those with the greatest p-value), resulting in only variables that significantly predicted habitat use (P 0.05). The remaining significant variables were used to develop a model of habitat use.

PAGE 22

14 Figure 2-1. Map of Santa Rosa Island, Eglin Air Force Base, Florida with location of six study sites (trapping grids) identified by black squares (). Figure 2-2. Black and white satellite imagery (IKONOS, May 2001) depicting an aerial view of Site E showing placement of the trapping grid among five landscape elements [dune fragment (found within the frontal zone), regenerating or remnant interior dunes, coastal grasslands and swales].

PAGE 23

15 Table 2-1. Eleven variables illustrating the variation of dune fragment isolation and connectivity among six study sites (trapping grids, see Figure 2-1). The variables in bold were used in regression analyses. Site Dune Fragment Area (ha) Dune Fragment Perimeter (m) Regenerating or Remnant Interior Dunes (%) Swale Habitat (%) Cover of food species (%) Cover of grass species (%) A 3.2 779.8 58.2 26.4 19.2 28.3 B 2.0 585.8 47.3 17.6 21.7 26.5 C 1.4 337.6 27.5 40.7 11.5 22.5 D 0.8 320.5 40.7 42.9 13.0 15.3 E 0.6 229.2 38.5 1.1 9.4 12.3 F 0.2 106.0 18.7 12.1 6.4 7.4 Site Cover of herbaceous species (%) Species richness (count) Total Cover of vegetation (%) Mean Gap or Interdunal Distance (cm) Connectivity value (%) A 3.6 39 50.7 54.4 55.9 B 2.9 29 42.6 79.4 43.6 C 4.6 35 39.6 67.6 29.9 D 0.4 17 23.2 123.4 44.3 E 0.7 19 19.5 185.2 40.0 F 0.5 24 13.5 166.4 12.9

PAGE 24

16 Joined Unjoined Figure 2-3. Sample 5x5 grid demonstrates the calculation of the connectivity value for study sites. Each grid cell represents a trap location. Shaded areas are habitat (dune fragment, regenerating or remnant interior dune) and non-shaded areas are non-habitat (interdunal gap or swale). Connectivity value is calculated as the proportion of actual “joinings” to potential “joinings”. The arrows demonstrate joined and unjoined cells for an example trap location. Connectivity value for this grid with 22 actual joinings and 72 potential “joinings” would be calculated as 22/72 = 0.31

PAGE 25

RESULTS Beach mouse movement was strongly associated with dune fragment isolation and development. Movement (MSD) was significantly influenced by the amount of regenerating or remnant interior dune habitat surrounding the dune fragment (F = 11.78, df = 1, P = 0.003, r 2 = 0.40), indicating that dune development affects the amount and distances of beach mouse movement. MSD was reduced in habitat patches with increased interior dune development (Figure 3-1A). Also, beach mouse movement demonstrated a general trend of decreasing with decreased dune fragment isolation (Figure 3-2). These results suggest one of two theories: greater dune development increasingly limits beach mouse movement; or, increased interior dune development reduces the need for increased beach mouse travel. Three additional habitat variables were eliminated from the regression procedure, perimeter of dune fragment, area of the dune fragment, and connectivity value. Each of these correlations was highly positive (P < 0.05) with movement, and highly correlated with each other. Autocorrelation between these variables is common in fragmented habitats. These three factors exhibit a trend relational to movement (Figure 3-1B, C, & D), and thus, were most likely removed from the model because of autocorrelations between variables (Table 3-1). A threshold level of connectivity effecting movement was not determined with the sites included in this study. As values of connectivity decreased, movement increased linearly. More sites with increasingly isolated dune fragments would need to be examined 17

PAGE 26

18 to determine the connectivity threshold. Although, it may be difficult to see a threshold in a natural landscape such as the one examined due to a lack of a stable population within these habitat patches. Habitat structure significantly influences habitat use, as demonstrated by a multi-factorial model (Table 3-2). The following model explains habitat use (HAB), where agdIHABse0003.0013.012.0 I is a control variable dependent on season (s) and landscape element (e) (Table 3-3), d is distance from vegetation patch, g is the percent cover of grasses, and a is average gap or interdunal distance. The model demonstrates that habitat use is affected by vegetation and landscape structure, and beach mouse use increased with increased cover of dune grasses, decreased distance from a vegetation patch, and decreased average gap or interdunal distance. Beach mice selected the dune fragment landscape element preferentially over other landscape elements throughout the year (Figure 3-3). The preference for mature, developed dunes suggests that these dunes are important to the ecology of beach mice populations. However in the fall, mean habitat use (HAB) for dune fragment is reduced and is comparable to usage in other landscape elements, particularly regenerating or remnant interior dunes. These results suggest that dune fragment habitat is particularly important during the winter, spring and summer months, but less important during the fall season. Beach mice use of swale landscape element suggests this habitat may also be important, particularly in winter and spring when swales are often dry. The increased late summer and fall precipitation of this region saturates swales, effectively reducing beach

PAGE 27

19 mouse use. These results indicate that connectivity may also be influenced seasonally by swales and suggest a need to recalculate connectivity values. In the winter and spring months, swale habitat could be considered habitat, whereas in late summer and fall, non-habitat. To test the effect this may have on movement, connectivity was recalculated with swale habitat as “joined” with other landscape elements. The resulting regression demonstrates the new connectivity value does not significantly influence movement (F = 0, df = 1, P = 0.98, r 2 = 0.0001). Although the inclusion of swale as habitat was warranted due to habitat use, connectivity of these landscapes does not influence movement. These connectivity values do not take into account the influence of edge nor does it allow for some landscape elements to be ranked based on use. For example, the dune fragment is used more than any other landscape element, so joined dune fragment habitat may have a value of 1, whereas joined swale habitat may have a lower value (ie. 0.6). Also, joined swale-dune fragment habitat could also take on a lower than 1 value, due to some effect of edge. This more complex value would require further calculation and calibration, but could be used to define a more complex calculation of connectivity based on true use of the landscape. These data could be obtained by examining trapping results or movement patterns of individual mice. Table 3-1. Pearson correlation coefficients among habitat patch variables. All variables are closely related with P < 0.001. Area of Dune Fragment Perimeter of Dune Fragment Amount of Interior Dune Development Perimeter of Dune Fragment 0.98 • • Amount of Interior Dune Development 0.80 0.90 • Connectivity Value 0.72 0.82 0.96

PAGE 28

20 Table 3-2. Results of the multivariate Poisson regression procedure testing for the effect of habitat structure on number of first captures for individual trap locations. Effect df 2 P Season (trapping session) 3 52.38 <0.0001 Landscape Element 3 77.86 <0.0001 Distance to Vegetation Patch (m) 1 11.53 0.0007 Connectivity Value 1 5.14 0.02 Percent Cover of Grasses (%) 1 53.46 <0.0001 Average gap or interdunal distance (cm) 1 4.29 0.038 Table 3-3. The control variable (I se ) for habitat use model, varying by season of trapping session and landscape element. Dune Fragment Regenerating or Remnant Interior Dune Interdunal Gap Swale Winter -0.41 -1.09 -1.01 -1.51 Spring -0.12 -0.80 -0.71 -1.22 Summer -0.67 -1.36 -1.27 -1.78 Fall -0.72 -1.41 -1.32 -1.83

PAGE 29

A. y = -2578x + 2109.3R2 = 0.4021020040060080010001200140016001800200000.10.20.30.40.50.60.7Proportion of secondary dune development (%)Mean squared distance (m2) B. y = -282.8x + 1503R2 = 0.3384020040060080010001200140016001800200000.511.522.533.5Area of dune fragment (Ha)Mean squared distance (m2) C. y = -1.3732x + 1662.9R2 = 0.371302004006008001000120014001600180020000100200300400500600700800900Perimeter of the dune fragment (m)Mean sqaure distance (m2) D. y = -2666.1x + 2142.7R2 = 0.3724020040060080010001200140016001800200000.10.20.30.40.50.6Connectivity valueMean sqaured distance (m2) 21 Figure 3-1. Linear regression of mean square movement (MSD) and habitat patch variables: amount of regenerating dune habitat (A), area of dune fragment (B), perimeter of dune fragment (C), and connectivity value (D). Actual values are the diamond points. The regression line equation and r-square value are displayed on the graph. All relationships were significant at P < 0.05.

PAGE 30

22 0200400600800100012001400160018002000Lowest (A)BCDEHighest (F)Level of dune fragment isolation (Site)Mean MSD (m2) Figure 3-2. Mean movement (SE) for six levels of dune fragment isolation and fragmentation. Sites were ranked based on eleven variables of dune fragment isolation and fragmentation, with Site A having the most isolated dune fragment and Site F the least isolated. Movement (MSD) was averaged for all four trapping sessions (n = 4, except for Site E and F, where n = 2). 00.20.40.60.811.21.4WinterSpringSummerFallMean number of first captures Dune Fragment Regenerating or Remnant Interior Dunes Interdunal Gaps Swale Figure 3-3. Mean number of first captures (SE) per trap. Data are averaged for each landscape element (dune fragment, regenerating or remnant interior dune, interdunal gap, and swale) and season (winter, spring, summer, and fall). Higher values represent an increased capture history for the trap and, therefore, a higher preference for the landscape element.

PAGE 31

23 Table 3-4. Due to a greater than expected use of swale habitat, connectivity value was recalculated for the sites within the Winter and Spring trapping sessions to include swale landscape element as “joined” with other landscape elements. Site Original Connectivity Value New Connectivity Value A 55.9 100 B 43.6 71.9 C 29.9 82.4 D 44.3 42.8 E 40.0 28.1 y = -22.713x + 895.57R2 = 0.0001020040060080010001200140016001800200000.20.40.60.811.2New Connectivity ValueMean sqaured distance (m2) Figure 3-4. Linear regression of mean square movement (MSD) and the new connectivity value that incorporates swale as “joined” habitat. The values displayed are for Sites A, B C, D, and E during the Winter and Spring trapping sessions only. Actual values are the diamond points. The regression line equation and r-square value are displayed on the graph. The relationship was not significant (P = 0.97).

PAGE 32

DISCUSSION The conventional belief that increased interdunal gaps and dune isolation reduce beach mouse movement was not reinforced by this study. My results indicate that beach mice are moving further with increased dune fragment isolation and greater interdunal gap distances at the habitat patch scale. The amount of regenerating or secondary dune habitat associated with a dune fragment and landscape connectivity had a significant negative relationship to beach mouse movement. These results suggest that beach mice respond to loss of connectivity by increasing distance moved. Evidence of small mammals responding to fragmentation with increased movement has been reported in other studies, suggesting that small mammals are more capable of moving across matrix habitat than assumed. Following a hurricane, one radio-collared beach mouse was documented moving over 100 m across open habitat during nightly foraging excursions (Swilling et al. 1998). Diffendorfer et al. (1995) found that three small mammal species either moved further distances or moved less often in fragmented habitat than they did in continuous habitat. Wauters et al. (1994) examined space use responses of red squirrels (Sciurus vulgaris) to habitat fragmentation. Space use and home range increased in more fragmented habitat. For small habitat patches, space use and home range were influenced positively by the presence of corridors (reduced isolation). Although landscape variables suggest beach mice movement increases in more isolated habitat, microhabitat selection analyses indicate a preference for vegetation cover 24

PAGE 33

25 and connectivity. Beach mice significantly selected habitat for increased percent cover of vegetation, decreased average gap or interdunal distance, and decreased distance from a vegetation patch. This is similar to observations by Bird (2003), who found vegetation cover influenced foraging behavior of beach mice, with increased residence time and foraging in sites with increased cover. These results suggest that increased movement in more isolated habitat may be due to an increase in distances between resources and not due to preference for gap and interdunal habitat. Other studies on habitat use suggest that beach mice prefer dune habitat with sparse vegetation and low cover (Blair 1951, Pournelle and Barrington 1955, Ivey 1959, Humphrey and Barbour 1981), contrary to my results of a preference toward increased cover. These studies considered vegetation cover across all dune habitats including areas with dense cover of herbaceous and woody species found in late-successional habitats. My study was conducted in habitat recovering from hurricane damage and does not contradict these findings. During this period of early recovery, vegetative cover was less than 50% and woody scrub species cover was less than 2%. It is apparent that beach mice prefer early-successional dune habitat within the frontal zone. Beach mice will also use interior and late-successional scrub habitat, but to a lesser degree. No biological evidence suggests either habitat is less advantageous, either to fitness or survival (Swilling 2000). But in habitats severely disturbed by hurricane and storm impacts, beach mice will travel further in order to obtain necessary resources. Providing vegetative cover and foraging plants, through habitat restoration, will reduce movement distances and potentially increase fitness and population reestablishment following catastrophic storm events.

PAGE 34

26 Beach mice use of swale habitat was nearly equal to or greater than the use of interior dune habitat during dry seasons. These results were surprising, since previous studies indicate swale habitat as “inadequate” and not used by beach mice (Blair 1951). Ephemerally wet swale habitats have a high concentration of insects relative to other landscape elements (personal observation) and insects are a dominant food source for beach mice during all seasons (Moyers 1996). Swale may be as important to beach mice as interior dunes in providing foraging resources. A more intensive study on the impact of swale habitat, insect population, and insect consumption on beach mice populations are necessary. Simulation modeling suggests that critical thresholds exist for species in fragmented landscapes (With and Crist 1995). A critical threshold was not found in this study. Beach mice may be adapted to the disturbance regime of a naturally dynamic environment and can increase movement in response to level of fragmentation and habitat patch isolation found in this system. As theory suggests, there may still be a point where movement is reduced under severe patch isolation, but these conditions were not reached within the sites chosen for this study. It is important to point out that the possibilities of detecting significant effects depends on sample size, the magnitude of the effect, and the variables measured. With a sample size of six (for some trapping sessions only four or five), it would be plausible that more sampling is needed to determine a threshold, should one exist. Particularly when comparing continuous dune habitat with severely isolated dune fragments. Selected sites represented fragmentation and isolation available within the study area. The

PAGE 35

27 available extremes (least isolated site vs. the most isolated) may not adequately represent the extremes necessary to detect a critical threshold at this scale. Habitat fragmentation research has focused on the influence of patch isolation and habitat-loss at the kilometers-wide scale, but organism’s function across a wide range of scales and can demonstrate vulnerability to fragmentation at these scales. Fragmentation is therefore not scale-limited (Lord and Norton 1990), and the concept of fragmentation should be examined at the scale most likely to influence the species in question. Beach mice demonstrated an increase in movement as habitat isolation increased. Further investigation at a finer scale, demonstrates a preference for vegetation cover and connectivity. A removal of these microhabitat variables, as is characteristic of isolated habitats, may have negative consequences for beach mice. Understanding the habitat-specific movement abilities of organisms in fragmented habitat is critical for addressing management and conservation concerns (Merriam 1991). These results have important implications for the conservation of the Santa Rosa beach mouse, and potentially all Gulf coast beach mouse subspecies. The ability of animals to persist in dynamic habitats is directly related to mobility (Ims 1995). Since beach mice may be required to move further in fragmented habitat and the overall amount of preferred habitat for beach mice is reduced, habitat fragmentation and isolation may affect beach mice survival and longevity. Increased movement distances can influence populations by increasing risk to predation. Also, additional energy spent moving to acquire resources may result in a tradeoff with energy available for other functions, such as dispersal, finding suitable mates, reproduction, etc. A reduction in these functions could affect survival and longevity of populations within fragmented habitat. Further

PAGE 36

28 investigation is warranted on the influence of habitat isolation and loss of connectivity on survival and reproduction of beach mouse populations. In addition, the effect of swale habitat and insects as foraging resources requires additional study to determine its influence on populations in recovering habitat.

PAGE 37

APPENDIX RESEARCH DATA Table A-1. List of food plants used in calculating percent cover of food species (Moyers 1996). Common Name Species Name Season Cakile Cakile spp. * Beach pea Galactia spp. Spring/Fall Seaside pennywort Hydrocotyle bonariensis Spring Inkberry Ilex glabra * Yaupon holly Ilex vomitoria Spring Seashore elder Iva imbricata Spring/Fall Evening primrose Oenothera humifusa Winter/Spring/Summer/Fall Panicgrass Panicum amarum * Ground cherry Physalis angustifolia Spring/Summer Bluestem Schizachyrium maritimum Winter/Fall Greenbriar Smilax spp. Spring Sea oats Uniola paniculata Winter/Summer/Fall * Indicates a species that was not listed as a food plant by Moyers 1996, but field evidence suggests the species may be an important food plant for Santa Rosa beach mouse. Moyers, J. E. 1996. Food habitas of the Gulf Coast subspecies of beach mice (Peromyscus polionotus spp.). M.Sc. Thesis, Auburn University, Auburn. 29

PAGE 38

30 Table A-2. Pearson correlation coefficients for variables used in landscape analysis. Shaded boxes have a Pearson correlation coefficient > 0.70 PERM DUNE SWAL FOOD GRAS HERB SPPR PCOV AGAP CVAL AREA 0.98 0.80 0.04 0.81 0.91 0.46 0.81 0.95 -0.83 0.78 PERM 0.90 0.02 0.90 0.85 0.36 0.68 0.90 -0.78 0.87 DUNE -0.07 0.83 0.54 -0.01 0.23 0.65 -0.26 0.97 SWAL -0.03 -0.81 0.14 0.09 0.17 -0.46 0.28 FOOD 0.73 0.07 0.28 0.61 -0.37 0.81 GRAS 0.71 0.86 0.98 -0.90 0.36 HERB 0.92 0.78 -0.92 0.04 SPPR 0.93 -0.99 0.32 PCOV -0.96 0.69 AGAPS -0.62 AREA Area of Dune Fragment (ha) PERM Perimeter of Dune Fragment (m) DUNE Amount of Regenerating or Remnant Interior Dune Development (%) SWAL Amount of Swale Habitat (%) FOOD Total Percent Cover of Food Species (%) GRAS Total Percent Cover of Grass Species (%) HERB Total Percent Cover of Herbaceous Species (%) SPPR Species Richness PCOV Total Percent Cover of Vegetation (%) AGAPS Average Gap or Interdunal Distance (cm) CVAL Connectivity Value

PAGE 39

Table A-3. Pearson correlation coefficients for microhabitat variables used in habitat selection analyses. 31 PCOV DIST CVAL SPPR FOOD GRASS HERB LARG AGAP LAND -0.07 0.26 -0.16 -0.07 -0.50 -0.16 0.03 0.24 0.28 PCOV -0.39 -0.04 0.69 0.59 0.92 0.52 -0.76 -0.57 DIST 0.11 -0.39 -0.31 -0.38 -0.17 0.58 0.63 CVAL -0.04 0.10 -0.02 -0.01 0.10 0.13 SPPR 0.26 0.57 0.49 -0.73 -0.64 FOOD 0.57 0.49 -0.54 -0.44 GRASS 0.52 -0.73 -0.56 HERB -0.34 -0.24 LARG 0.86 *High correlation values (>0.70) were removed from multiple Poisson regression analyses and are shown in bold LAND Landscape Element PCOV Total Percent Cover of Vegetation DIST Distance to Nearest Vegetation Patch CVAL Connectivity Value SPPR Species Richness FOOD Total Percent Cover of Food Species GRASS Total Percent Cover of Grass Species HERB Total Percent Cover of Herbaceous or Non-grass Species LARG Largest Gap or Interdunal Distance AGAP Average Gap or Interdunal Distance

PAGE 40

A. y = -282.8x + 1503R2 = 0.3384020040060080010001200140016001800200000.511.522.533.5Area of dune fragment (Ha)Mean squared distance (m2) B. y = -5146.1x + 1814.2R2 = 0.2809020040060080010001200140016001800200000.050.10.150.20.25Total percent cover of food vegetationMean sqaure distance (m2) C. y = -3755.5x + 1832.9R2 = 0.2665020040060080010001200140016001800200000.050.10.150.20.250.3Total percent cover of grass species Mean square distance (m2) D. y = 3E+06x2 96667x + 1493R2 = 0.3104020040060080010001200140016001800200000.0050.010.0150.020.0250.030.0350.04Total percent cover of herbaceous speciesMean square distance (m2) E. y = -19.066x + 1581.2R2 = 0.11420200400600800100012001400160018002000051015202530354045Species richnessMean square distance (m2) 32 Figure A-1. Linear regression of mean square movement (MSD) and habitat patch variables: amount of regenerating dune habitat (A), amount of swale habitat (B), total percent cover of food species (C), total percent cover of grasses, (D) total percent cover of herbaceous species, and (E) species richness. Actual values are the diamond points. The regression line equation and r-square value are displayed on the graph.

PAGE 41

33 Table A-4. Number of individual mice captured per site, by trapping session and gender (M = Male, F = Female). Site Sex Winter Spring Summer Fall Total A M 15 15 11 5 46 F 13 13 8 6 40 Total 28 28 19 11 86 B M 12 9 9 13 43 F 11 6 5 7 29 Total 23 15 14 20 72 C M 10 11 6 6 33 F 15 12 4 4 35 Total 25 23 10 10 68 D M 8 13 6 4 31 F 12 15 7 2 36 Total 20 28 13 6 67 E M • 12 7 8 27 F • 9 5 3 17 Total • 21 12 11 44 F M • • 3 5 8 F • • 4 4 8 Total • • 7 9 16

PAGE 42

LIST OF REFERENCES Barnum, S. A., C. J. Manville, J. R. Tester, and W. J. Carmen. 1992. Path selection by Peromyscus leucopus in the presence and absence of vegetation cover. Journal of Mammalogy 73:797-801. Bird, B. L. 2003. Effects of predatory risk, vegetation structure, and artificial light on the foraging behavior of beach mice. M.Sc. University of Florida, Gainesville. Blair, W. F. 1951. Population structure, social behavior, and environmental relations in a natural population of the beach mouse (Peromyscus polionotus leucocephalus). Contributions from the Laboratory of Vertebrate Biology 48:1-47. Bowne, D. R., J. D. Peles, and G. W. Barrett. 1999. Effects of landscape spatial structure on movement patterns of the hispid cotton rat (Sigmodon hispidus). Landscape Ecology 14:53-65. Brooker, L., M. Brooker, and P. Cale. 1999. Animal dispersal in fragmented habitat: measuring habitat connectivity, corridor use, and dispersal mortality. Conservation Ecology 3:4. Chafin, L., C. Kindell, B. Herring, C. Nordman, J. Jensen, and A. Schotz. 1997. Natural community survey of Eglin Air Force Base, 1993-1996: Final Report. Florida Natural Areas Inventory, Tallahassee, FL. Collinge, S. K. 2000. Effects of grassland fragmentation on insect species loss, colonization, and movement patterns. Ecology 81:2211-2226. Delin, A. E., and H. Andrn. 1999. Effects of habitat fragmentation on the Eurasian red squirrel (Sciurus vulgaris). Landscape Ecology 14:67-72. Diffendorfer, J. E., M. S. Gaines, and R. D. Holt. 1995. Habitat fragmentation and movements of three small mammals (Sigmodon, Microtus, and Peromyscus). Ecology 76:827-839. Fahrig, L., and G. Merriam. 1985. Habitat patch connectivity and population survival. Ecology 66:1762-1768. Fahrig, L., and G. Merriam. 1994. Conservation of fragmented populations. Conservation Biology 8:50-59. 34

PAGE 43

35 Foster, J., and M. S. Gaines. 1991. The effects of a successional habitat mosaic on a small mammal community. Ecology 72:1358-1373. Gore, J. A., and T. L. Schaefer. 1993a. Distribution and conservation of the Santa Rosa beach mouse. Proceedings Annual Conference Southeast Association of Fish and Wildlife Agencies 47:378-385. Gore, J. A., and T. L. Schaefer. 1993b. Final performance report: Santa Rosa beach mouse survey. Florida Game and Fresh Water Fish Commission, Panama City. Holler, N. R. 1992a. Choctawhatchee beach mouse. Pages 76-96 in S. R. Humphrey, editor. Rare and endangered biota of Florida. University Press of Florida, Gainesville. Holler, N. R. 1992b. Perdido Key beach mouse. Pages 102-109 in S. R. Humphrey, editor. Rare and endangered biota of Florida. University Press of Florida, Gainesville. Humphrey, S. R., and D. B. Barbour. 1981. Status and Habitat of 3 Subspecies of Peromyscus-Polionotus in Florida. Journal of Mammalogy 62:840-844. Ims, R. A. 1995. Movement patterns related to spatial structures. Pages 85-109 in L. Hansson, L. Fahrig, and G. Merriam, editors. Mosaic Landscapes and Ecological Processes. Chapman & Hall, London. Ivey, R. D. 1959. Life history notes on three mice from the Florida East Coast. Journal of Mammalogy 30:162-164. Kaufman, D. W., S. K. Peterson, R. Fristik, and G. A. Kaufman. 1983. Effect of microhabitat features on habitat use by Peromyscus leucopus. American Midland Naturalist 110:177-185. Kincaid, W. B., and G. N. Cameron. 1985. Interactions of Cotton rats with a patchy environment: dietary responses and habitat selection. Ecology 66:1769-1783. Kozakiewiez, M. 1993. Habitat isolation and ecological barriers: the effect on small mammal populations and communities. Acta Theriologica 38:1-30. Lord, J. M., and D. A. Norton. 1990. Scale and the spatial concept of fragmentation. Conservation Biology 4:197-202. Lynn, W. J. 2000. Social organization and burrow-site selection of the Alabama beach mouse (Peromyscus polionotus ammobates). M. Sc. Thesis. Auburn University, Auburn. McGarigal, K. 1995. FRAGSTATS: Spatial pattern analysis program for quantifying landscape structure. PNW-351, USDA For Serv Gen Tech Rep.

PAGE 44

36 McMillan, B. R., and D. W. Kaufman. 1995. Travel path characteristics for free-living White-footed mice (Peromyscus leucopus). Canadian Journal of Zoology 73:1474-1478. Merriam, G. 1984. Connectivity: A fundamental ecological characteristic of landscape pattern. Pages 5-15 in J. Brandt and P. Agger, editors. Methodology in landscape ecological planning, Roskilde University Centre, Denmark. Merriam, G. 1991. Corridors and connectivity: animal populations in heterogenous environments. Pages 133-142 in D. A. Sounders and R. J. Hobbs, editors. Nature conservation 2: the role of corridors. Surrey Beatty & Sons. Moyers, J. E. 1996. Food habits of the Gulf Coast subspecies of beach mice (Peromyscus polionotus spp.). M. Sc. Thesis. Auburn University, Auburn. Pither, J., and P. D. Taylor. 1998. An experimental assessment of landscape connectivity. Oikos 83:166-174. Planz, J. V., and J. Kirkland, G. L. 1992. Use of woody ground litter as a substrate for travel by the White-footed mouse, Peromyscus leucopus. Canadian Field Naturalist 106:118-121. Pournelle, G. H., and B. A. Barrington. 1955. Notes on mammals of Anastasia Island, St. Johns County Florida. Journal of Mammalogy 334:133-135. SAS Institute, I. 1999. SAS/STAT. in. SAS Institute, Inc., Cary, North Carolina. Saunders, D. A., R. J. Hobbs, and C. R. Margules. 1991. Biological conservation of ecosystem fragmentation: a review. Conservation Biology 5:18-32. Slade, N. A., and L. A. Russel. 1998. Distances as indices to movements and home-range size from trapping records of small mammals. Journal of Mammalogy 79:346-351. Slade, N. A., and R. K. Swihart. 1983. Home range indices for the hispid cotton rat (Sigmodon hispidus) in northeastern Kansas. Journal of Mammalogy 64:580-590. Sneckenberger, S. I. 2001. Factors influencing habitat use by the Alabama beach mouse (Peromyscus polionotus ammobates). M. Sc. Thesis. Auburn University, Auburn. Stamps, J. A., M. Buechner, and V. V. Krishnan. 1987. The effects of edge permeability and habitat geometry on emigration from patches of habitat. The American Naturalist 129:533-552. Swilling, J., W. R. 2000. Ecological dynamics of the endangered Alabama beach mouse (Peromyscus polionotus ammobates). M. Sc. Thesis. Auburn University, Auburn.

PAGE 45

37 Swilling, J., W. R., M. C. Wooten, N. R. Holler, and W. J. Lynn. 1998. Population dynamics of Alabama beach mice (Peromyscus polionotus ammobates) following Hurricane Opal. American Midland Naturalist 140:287-298. Szacki, J., and A. Liro. 1991. Movements of small mammals in the heterogenous landscape. Landscape Ecology 5:219-224. Taylor, P. D. 1993. Connectivity is a vital element of landscape structure. Oikos 68:571-573. Wauters, L., P. Casale, and A. A. Dhondt. 1994. Space use and dispersal of Red squirrels in fragmented habitats. Oikos 69:140-146. Wegner, J., and G. Merriam. 1990. Use of spatial elements in a farmland mosaic by a woodland rodent. Biological Conservation 54:236-276. Wiens, J. A., N. C. Stenseth, B. Van Horne, and R. A. Ims. 1993. Ecological mechanisms and landscape ecology. Oikos 66:369-380. With, K. A., and T. O. Crist. 1995. Critical thresholds in species responses to landscape structure. Ecology 76:2446-2459.

PAGE 46

BIOGRAPHICAL SKETCH Kathryn Enga Louise Smith, daughter of James H. and Darlene C. Smith, was born on October 12, 1975, in Robinsdale, Minnesota. She graduated from Champlin Park High School in Brooklyn Park, Minnesota, in June of 1994. She attended North Hennepin Community College. During her studies, she worked in a chemistry lab as a chemistry lab assistant and tutor. She graduated with honors with her Associate of Arts degree in December 1995. She transferred to the University of Minnesota, Twin Cities, in January 1996. As a student, she worked as an intern for Dr. David Tilman at the Cedar Creek Natural History Area in Bethel, Minnesota. She also worked in a plant pathology lab at the University of Minnesota, as an intern at Springbrook Nature Center for the City of Fridley, and as a student worker for the Minnesota Department of Natural Resources, Division of Minerals. She received her Bachelor of Science degree in natural resources and environmental studies with distinction in June 1998. She continued to work for the Minnesota Department of Natural Resources until deciding to attend graduate school to further her education in ecology. She entered graduate school at the University of Florida, Gainesville, Florida in August 1999. She worked with Drs. Deborah Miller and Mack Thetford and held an assistantship at the West Florida Research and Education Center in Milton, Florida, where she assisted with dune restoration research and wildlife courses and began graduate thesis research. 38