Citation
The effect of mowed paths in a grassland habitat on the dispersion of the prairie vole (microtus ochrogaster)

Material Information

Title:
The effect of mowed paths in a grassland habitat on the dispersion of the prairie vole (microtus ochrogaster)
Creator:
Dickerson, Elizabeth Ann
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
52 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Prairie vole -- Habitat -- Colorado -- Jefferson County ( lcsh )
Ecological disturbances ( lcsh )
Environmental impact analysis ( lcsh )
Prairie vole -- Effect of habitat modification on -- Colorado ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 46-52).
Thesis:
Biology
General Note:
Department of Integrative Biology
Statement of Responsibility:
by Elizabeth Ann Dickerson.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
45141388 ( OCLC )
ocm45141388
Classification:
LD1190.L45 2000m .D55 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

THE EFFECT OF MOWED PATHS IN A GRASSLAND HABITAT ON THE DISPERSION OF THE PRAIRIE VOLE (MICROTUS OCHROGASTER) by Elizabeth Ann Dickerson B.S., Metropolitan State College of Denver, 1984 B.S., University of Colorado, 1986 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Master of Arts Biology 2000 '' .. I ; i, L . ._ ....

PAGE 2

This thesis for the Master of Arts degree by Elizabeth Ann Dickerson has been approved by Cheri A. Jones Diana F. Tomback Teresa Audesirk C) /;o/ ;1 cs ob ) Date

PAGE 3

Dickerson, Elizabeth Ann (M.A., Biology) The Effect of Mowed Paths in Grassland Habitats on the Dispersion of the Prairie Vole (Microtus ochrogaster) Thesis directed by Adjunct Professor Cheri A. Jones ABSTRACT I studied the dispersion of Prairie Voles (Microtus ochrogaster) relative to a 2 m-wide mowed path in a 4.1-ha grassland at the Two Ponds National Wildlife Refuge, Jefferson County, Colorado. The effect of the path on vole home range size, maximum distance traveled, frequency of path crossing, and grassland vegetation structure were studied. Animals were livetrapped May through October 1999 in three grids one control grid was situated in contiguous habitat and two test grids straddled a previously established path. Each grid contained 60 traps, placed 5 m apart, arranged in six transects of ten traps per transect. Ground cover, litter depth, vegetation cover, and vegetation height were measured in a 25-m2 area that centered on each trap. The distances between trap to path and landscape features open water and different vegetation community were recorded. All captures were associated with 94.2 5.5% ground cover, 2.50 1.30 em litter depth, and 55.57 15.55 em vegetation height; noncapture sites differed from capture sites in at least one characteristic. In Ill

PAGE 4

between trap to path and landscape features were recorded. All captures were associated with 94.2 5.5% ground cover, 2.50 1.30 em litter depth, and 55.57 15.55 em vegetation height; non-capture sites differed from capture sites in at least one characteristic. In test grids, the number of captures, litter depth, and vegetation height were positively correlated with distance from the path. Low capture number and poor habitat quality were associated with a 4-m margin on both sides of the path, indicating an edge effect. No home ranges included the path or were closer than 4 m from the path's edge. No individuals switched sides of the path. Habitat structure and capture patterns were spatially homogeneous in the control grid. Population density increased as the area of contiguous preferred habitat increased. Mammalian predators used the path as a corridor. In general, the path and disturbed margins influenced the dispersion of Prairie Voles by facilitating predator access, reducing the available habitat by 21%, increasing the distance between suitable patches, and contributing to a reduction of carrying capacity and population size. Dispersion across the path could be inhibited due to a lack of suitable cover. This abstract accurately represents the content of the candidate's thesis. recommend its publication. IV

PAGE 5

DEDICATION I dedicate this thesis to Carol, Molly, Mica, and Roz.

PAGE 6

ACKNOWLEDGEMENT Thanks to David Jamiel, Ray Rauch, and John Comely of the U.S. Fish and Wildlife Service for generous funding, supervision, and inspiration for this project. Thanks to the Zoology Department at the Denver Museum of Natural History for use of the Mammal Collection and equipment. I wish to express appreciation to my committee members Cheri Jones, Diana Tomback, and Terri Audesirk-for their time, effort, and invaluable expertise Special thanks to Cheri Jones for her guidance and support.

PAGE 7

CONTENTS Figures .................................................................................................. ix Tables .................................................................................................... x Chapter 1. Introduction ........................................................................................ 1 2. Materials and Methods ...................................................................... 6 2. 1 Study Site ........................................................................................ 6 2.2 Field Methods ................................................................................. 8 2.3 Analytical Methods ........................................................................ 11 3. Results ............................................................................................ 15 3.1 Capture Summary ......................................................................... 15 3.2 Habitat Characteristics at Capture and Non-capture Sites ............................................................................................. 18 3.3 Spatial Relationship of Captures and Habitat Features ....................................................................................... 24 3.4 Influence of the Path on Vegetation ............... .............................. 28 3.5 Home Range and Movement Analysis .................... ..................... 28 4. Discussion ....................................................................................... 35 5. Conclusion ....................................................................................... 42 6. Management Recommendations ........... ........ ........ ....................... 43 VII

PAGE 8

References ........................................................................................... 46 Vlll

PAGE 9

FIGURES Figure 2. 1 Map of Two Ponds National Wildlife Refuge Environmental Education Area ........................................................ 7 2.2 Grid Trap Arrangement Model ......................................................... 8 2.3 Home Range by Inclusive Boundary Strip Method ......................... 13 3.1 Grid 1 (Test) Habitat Characteristics .............................................. 22 3.2 Grid 2 (Test) Habitat Characteristics .............................................. 22 3.3 Grid 3 (Control) Habitat Characteristics ......................................... 23 3.4 Grid 1 (Test) Habitat Characteristics and Capture Number Correlated with Distance from the Path ............................ 24 3.5 Grid 2 (Test) Habitat Characteristics and Capture Number Correlated with Distance from the Path ............... ...... .... 25 3.6 Grid 3 (Control) Habitat Characteristics and Capture Number Correlated with Distance from the Path ........................... 26 3. 7 Grid 1 (Test) Prairie Vole Home Ranges ....................................... 32 3.8 Grid 2 (Test) Prairie Vole Home Ranges ....................................... 33 3. 9 Grid 3 (Control) Prairie Vole Home Ranges ................................... 34 IX

PAGE 10

TABLES Table 3.1 Trap Summary ............................................................................ ... 15 3.2 Prairie Vole Population Density ..................................................... 17 3.3 Habitat Characteristics at Capture and Non-capture Sites .............................................................................................. 20 3.4 Mean Values for Significant Habitat Characteristics at Capture and Non-capture Sites ..................................................... 21 3.5 Correlation of Vegetation Structure, Capture Number, and Distance from the Path ........................................................... 27 3.6 Comparison of Home Range Size and Maximum Distance Traveled .......................................................................... 29 X

PAGE 11

1. Introduction Fragmentation of wildlife habitats, or separation of contiguous habitat into smaller patches, is increasing through human activities (Yahner 1988; Saunders et al. 1993; Nassauer 1995). All types of development, clearing of land for agricultural use, resource extraction, and building and use of roads decrease the area of suitable habitat and increase edge (Lynch and Whigman 1984; Bierregaard et al. 1992; Szacki et al. 1993; Nupp and Swihart 1996; Villard et al. 1999). Organisms have developed life histories and physiological adaptations to cope with natural or inherent fragmentation produced by variations in climate, topography, soil type, and vegetation communities (Curtis 1956; Thomas et al. 1979). A small mammal, the Meadow Vole (Microtus pennsylvanicus) for instance, is solitary and territorial as an adult. Female voles can more effectively defend a patch than contiguous habitat. Patchiness decreases competitive pressure, resulting in increased survival and production of offspring (Collins and Barrett 1997). However, patchiness, particularly patchiness created by human alteration of habitats, has negative impacts on most wildlife species. Patch edges, or junctions between two distinct landscape types, differ from

PAGE 12

interiors in their increased exposure to the desiccating and mechanical effects of wind, increased temperature, and extremes of light (Forman and God ron 1986; Kapos 1989; Laurance 1991 ). Small patch size makes residents more vulnerable to disease, predation, and demographic effects such as reduction of population densities (Wilcove 1985; Yahner 1988; Dooley and Bowers 1998). The process of recolonization of previously altered areas and dispersal from disturbed habitats may be disrupted due to an animal's inability to cross even the small area of unsuitable habitat presented by roads and highways (Getz et al. 1978; Swihart and Slade 1984; Bierregaard et al. 1992; Forman and Deblinger 2000). The habitat associated with road right-of-ways deteriorates due to increased edge, accentuating the barrier effect and reducing habitat (Haskell 2000). Research interest is also focusing on small-scale intrusions like footpaths and applying the knowledge gained to the design of parks and preserves (Mader 1984; Kasworm and Manley 1990). In this study, the effect of a mowed path on dispersion, or spatial distribution, of Prairie Voles (Microtus ochrogaster) was evaluated to see if small-scale fragmentation influences their life history. The project was conducted in a grassland community at the Two Ponds National Wildlife Refuge, Jefferson County, Colorado. 2

PAGE 13

Prairie Voles are rodents adapted to grasslands and found throughout the central U.S. and Canada (Stalling 1990, 1999). In Colorado they have been reported in four habitat types-grassy, grass-forb, shrubland, and riparian woodland habitats along the Front Range of the Rocky Mountains and the drainages of the South Platte and Arkansas rivers (Armstrong 1972; Carey 1978; Moulton et al. 1981; Reed and Choate 1988). Prairie Voles frequent the soil/vegetation interface, building ground nests under the protection of logs and other debris. In suitable soil, shallow burrows (5-30 em) and burrow systems are constructed (Fisher 1945; Carroll and Getz 1976; Stalling 1990). The presence of voles in dense vegetation is easily discerned by the existence of long, shallow depressions, cleared of vegetation, that function as highways or runways. Diet is comprised of a variety of herbaceous plants, grasses, roots, bark, arthropods, and carrion. Prairie Voles are preyed upon by a number of small to medium-sized carnivores including raptors, mammals, amphibians, and snakes (Martin 1956; Korschgen and Baskett 1963; Stalling 1990). Researchers have investigated the impact of anthropogenic habitat fragmentation on the activity of Prairie Voles. Getz et al. (1978) studied habitat preferences and the dispersal routes of voles within a network of interstate highways and country roads covering over 200 km2 Swihart and Slade (1984) examined Prairie Vole movements relative to a single-lane dirt 3

PAGE 14

road. Diffenderfer et al. (1995) experimentally created a patchwork of habitat islands, separated by at least 16 m, to determine dispersal patterns. The researchers discovered that Prairie Voles were not found near heavily traveled roads, and voles inhabited the tall grass margins along low trafficked country and dirt roads; animals were more likely to move away from a road or deteriorated habitat path than cross it. This study differs from previous investigations of corridor fragmentation on Prairie Voles in the scale of the disturbance; the path is only three years old, narrower than any type of road, contains no vehicle traffic, and has low pedestrian use. The effect of the path on the vegetation structure of vole habitat and vole use of habitat is emphasized rather than the animal's movements relative to the disturbance. The results can be applied to the design and maintenance of trail systems to alleviate their impact on small mammal populations. In order to assess the dispersion of voles relative to a small-scale disturbance, a 2-m wide grass corridor, I set up grids of small mammal live traps that straddled the path, and evaluated the resultant capture patterns. Three hypotheses were tested: (1) The frequency of capture and number of individuals captured will increase as the distance from the path increases; (2) Individuals do not include the path in their home ranges; (3) Voles cross the path less frequently than they move an equivalent distance in 4

PAGE 15

continuous habitat. To test these hypotheses, the habitat characteristics at capture and non-capture sites were compared to determine if variations in vegetation structure were associated with vole capture. The number of individuals that crossed the path, the distance traveled, and the frequency of crossings were observed. The home ranges of the animals were mapped when appropriate to see if they included the path as part of their range. 5

PAGE 16

2. Materials And Methods 2.1 Study Site Two Ponds National Wildlife Refuge is located in Jefferson County, Colorado (39 50'26" N, 105 06'15" W, elevation 2088 meters). The Refuge, managed by the U.S. Fish and Wildlife Service, encompasses 32.8 hectares of old pasture and farmland surrounded by suburban residential development. Two irrigation canals divide the property into three distinct sections. This study was performed in an 8.2-ha section (Figure 2. 1) containing 4. 1 ha of smooth brome (Bromus inermis) and alfalfa (Medicago sativa), three ponds surrounded by cattail (Typha latifolia) marsh, mature cottonwoods (Populus sp.), willows (Salix sp.), and Russian olive trees (Eiaeagnus angustifolia). A house and several small out buildings were removed from the property but the associated ornamental shade and fruit trees and shrubs still remain. The 8.2-ha area is managed for the protection of urban wildlife and migratory bird habitat. The area is surrounded by a 2-m high chain-link fence and locked gate. Access is limited to environmental education classes, maintenance, and scientific study. There are no vehicle roads in the study area. 6

PAGE 17

-..J TWO PONDS NATIONAL WILDLIFE REFUGE AREA _ ef N W*E s C Trapping grid -Ponds Tree / Mowed paths [-"] Cattail marsh Upland wood/shrub 0 Riparian woodland _-:-.J Grusland 200 0 200 Meters .. ------. -----. .. -----. -Figure 2.1 MAP OF STUDY AREA

PAGE 18

A footpath (approximately 2-m wide) was initially established in the grassy area in 1997 and is mowed monthly. 2.2 Field Methods Three trapping grids, each covering an area of 1500 m2 were established in the grassland (Figure 2.1 ). Each grid contained 60 Sherman live-traps (9 x 23 x 8 em) arranged in six parallel transects. Each transect contained 10 traps. Transects and traps were positioned 5 m apart (Figure 2.2). TRAP NUMBER .. w f}J -: .t :E r -, [2] '4' [6] 2: ;( [_9J !JoJ TRANSECT 'J> llJ [6J 171 }: '! r-l l!i [ij [3! i7' __ j [8] [_9] l}_i [2] : 3] (4] m w 9: m L., [3] :4J ] [6j [fJ [8j PATH FIGURE 2.2 GRID TRAP ARRANGEMENT MODEL A mowed 2-m wide path bisected the test grids (grids 1 and 2) between traps 5 and 6. The control grid (grid 3) did not include a path. Grid locations were selected by the presence or absence of a path, absence of irrigation ditches and bridges, and heterogeneity of aspect, topography, and 8

PAGE 19

vegetation community. Variations in habitat characteristics existed among grids, but the influence of the path was the primary concern. The data from test grids were compared to those from the control grid to determine whether the path had an influence on vegetation characteristics, home range size, and maximum distance traveled. The control grid was also used to detect the influence of the grid model itself on capture patterns. A trap session consisted of three consecutive nights. Prairie Voles are most active between sunset and midnight; traps were baited with oatmeal, set in the evening, and checked the following morning. Traps were closed during the day. Cracked corn and synthetic insulating fiber were added during cold and inclement weather. A bimonthly trap session, weather permitting, was conducted in each grid from May 1999 through October 1999. Captures were identified, weighed, sexed; reproductive and general condition was noted, and each animal was uniquely marked. A passive integrated transponder (PIT) (Destron-Fearing TX406L) encoding an alpha numeric identifier was implanted under the skin in the right shoulder area. Each captured animal was scanned with a portable reader (Destron-Fearing MPR) to detect the presence of a PIT. PIT tags are considered a humane alternative to mutilation marking and have been tested in laboratory and field applications (Fagerstone and Johns 1987; Hamner 1989; Elbin and Burger 1994; Carver et al. 1999). Incidental trap fatalities were submitted 9

PAGE 20

as voucher specimens to the Zoology Department at the Denver Museum of Natural History. The trap effective area, also called boundary strip, is defined as a square, surrounding and centered on the trap, having sides equal in length to the distance to the next trap (Stickel 1954). The vegetation characteristics for each trap's effective area were determined by taking a total of 50 measurements 70 em apart in a grid pattern. A pointed metal rod, 0.5 em diameter, was placed in the ground at sample locations (Owensby 1973). At the rod's point of contact, the presence or absence of ground cover (GCVR) and depth of litter (LOTH) in centimeters were measured using a meter stick. GCVR was considered any material other than bare ground; litter was composed of dead, flattened grass, leaves, and woody debris. The presence of live vegetation (VECVR), the live vegetation height (em) (VEGHGT), and identity of plant species within one centimeter of the rod were recorded. The distance of the trap to the path (DPTH), shortest distance to single woody shrub (DSRB), the shortest distance to dissimilar vegetation community or fence (DAL T), and distance to open water (DH20) were recorded. The trap area was sampled for these variables once between September 1 and October 30, 1999. 10

PAGE 21

2.3 Analytical Methods The number of animals was summed for each grid and reported as the number of total captures, number of individuals, and captures per trap night. Only data from captures of Prairie Voles were used in analyses. Population density was calculated by converting the number of individuals captured in each grid, which has an area of 0.15 ha, to the number of individuals in one hectare. GCVR and VECVR were recorded as present or absent, averaged for each trap area and expressed as a percent. The 50 LOTH and VEGHGT measurements were averaged for each trap area. The effective areas for traps 5 and 6 both included a 1 by 5 m section of the path. Nonparametric analyses were applied to vegetation characteristics because the data sets were seldom normally distributed. The Wilcoxon rank-sum test was used to determine whether variables at capture and non capture sites differed significantly within each grid and between grids. The habitat variables showing significant differences between capture and non capture sites were correlated with the number of captures (CAP#) and distance from the path (DPTH) within each grid using Spearman' rank correlation. Multiple captures of individuals and multiple captures at the 11

PAGE 22

same trap location were used in the correlation as indicators of habitat preference (Ribble and Samson 1987). Similar data from the test grids (1 and 2) were pooled and tested against the control (grid 3). DPTH was the distance from the trap to the center of the path in the transect. In grids 1 and 2 (test grids), the path was located between traps 5 and 6 in each transect (Figure 2.2). No path was present in grid 3 (control). In grid 3, the DPTH was the distance from the trap to a point midway between traps 5 and 6 in the transect. Chi-square with Yate's correction and the G-test were used to determine if captures were randomly distributed within grids. The home range was defined by Burt (1943) as the area in which an individual spends most of its time gathering food, mating, and caring for offspring. Home ranges of individuals 3 captures at more than one trap site were determined by the inclusive boundary strip method (Hayne 1949, 1950; Stickel1954; Shriner and Stacey 1991); the exterior corners of the capture site trap effective areas were connected to form a polygon (Figure 2.3). 12

PAGE 23

HOME RANGE BOUNDARY 0 01} :orD:[l_: AREA ; ....................... ; ...................... + ......... ) ....... l . ___ i D ..... ., .... ....... .J = CAPTURE LOCATION -= OUTLINE OF HOME RANGE Figure 2.3 HOME RANGE BY INCLUSIVE BOUNDARY STRIP METHOD The mean distance from the center of activity method (Hayne 1949, 1950) as applied by Slade and Swihart (1983) and Slade and Russell (1998), was used to determine maximum travel distances of voles with calculated home ranges. The center of activity is the geographic point represented by a two dimensional average, based on an x-y coordinate system, of capture locations. Mean square distance was calculated by the following formula: rl r-, Mean square distance =.L (x1-;)2 -t:L (Y1 Yl L"'-1 1,..-::j n-1 x 1 and y 1 are individual trap location coordinates, x and y are the mean of the x and y coordinates, and n is the number of captures. The mean 13

PAGE 24

square distance method is less sensitive to sample size than other methods (Diffenderfer 1995). Maximum distances traveled were normally distributed and therefore evaluated using the parametric F-test. All tests were considered significant at P s 0.05. 14

PAGE 25

3. Results 3.1 Capture Summary A trap-night is defined as one trap set for one night. Captures per trap-night are used to provide a comparison of capture rate between grids with different numbers of trap-nights (Table 3.1 ). GRID 1 GRID2 GRID3 TRAP NIGHTS 1259 1397 976 TOTAL CAPS PER TOTAL CAPS PER TOTAL CAPS PER CAPTURES TRAPNIGH CAPTURES TRAPNIGH CAPTURES TRAPNIGHT PRAIRIE VOLES 52[24] 0.041 123[49] 0.088 227[94] 0.233 MEADOW VOLES 0 0 0 0 3[2) 0.003 W. HARVEST MICE 5[4] 0.004 0 0 4[3] 0.004 HOUSE MICE 0 0 2[2) 0.001 6[4) 0.006 INDIVIDUALS = [ I CAPS = NUMBER OF CAPTURES TABLE 3.1 TRAP SUMMARY All small mammals live-trapped during the study In grids 1 and 3, 1440 trap-nights were attempted, but 1259 and 957, respectively, were completed. Grid 2 was trapped for 1620 nights, with 1397 completed. Rain, wind, and temperature changes had sprung traps in all grids. As I wore narrow paths in the grass of grid 3, Raccoons (Procyon /otor}, Red Foxes (Vulpes vulpes), and a Domestic Cat (Felis catus) raided and displaced traps more frequently. Paths were worn in Grids 1 and 2, but were disturbed less frequently than grid 3; traps adjacent to the path were 15

PAGE 26

sprung as often as traps at farther distances from the path. Grid 3 was located between two open water sources (an irrigation canal and pond), where Raccoons and Red Foxes hunted (personal observation). The worn path facilitated movement between the water sources. Small mammal diversity at the Refuge was similar to other areas in the Denver metropolitan area based on capture studies performed along the South Platte River corridor in 1998 and 1999 (personal observation and communication with C. A. Jones). In the grids, Prairie Vole populations out numbered House Mice (Mus musculus), Western Harvest Mice (Reithrodontomys mega/otis), and Meadow Voles (Microtus pennsylvanicus) by 12 to 1. Only Prairie Voles had home ranges within the grids and were recaptured at a rate higher rate than any other species. It is unlikely that the low population number of other rodent species captured affected the dispersion of Prairie Voles in the grassy areas. Prairie Voles are dominant over Meadow Voles in the laboratory and in grassy, xeric habitats (Findley 1954; Getz 1962). House mice are more likely associated with human structures, building rubble, and fencerows than open fields (Palmer and Fowler 1975). Western Harvest Mice are often sympatric with Prairie Voles, and no antagonistic behavior between the two species has been reported (Webster and Jones 1982). Population density increased from grid 1 to grid 3 (Table 3.2). 16

PAGE 27

TOTAL MALES FEMALES MALE: FEMALE ADULT SUBADULT JUVENILE ADULT SUBADULT JUVENILE ADULT SUBADULT JUVENILE GRID 1 160 73 7 33 27 7 13 2.70 1.00 2.54 GRID2 326 106 0 33 120 27 40 0.68 0.83 GRID3 626 213 33 100 206 27 47 1.03 1.22 2.13 Table 3.2 PRAIRIE VOLE POPULATION DENSITY (INDIVIDUALS/HECTARE) Population dynamics were similar in all grids; recapture rates were 42.33 3.21% (Student t-test, n = 3, d.f. = 2, t = 0.559, P (two tails) < 0.01 ); the absolute number of individuals with home ranges was proportional to density (25.72 4.53%) (Student t-test, n = 3, d. f. = 2, t = 0.311, P (two tails)< 0.01 ). Sex ratios of adults within populations approached parity as population density increased (Chi-square test, d. f. = 1; grid 1 X 2 = 21.160, P < 0.01; grid 2X2 = 0.867, P > 0.05; grid 3X2 = 0. 117, P > 0.05). The cause of a large discrepancy of males to females and juveniles to adults in grid 1 is unknown, but Gaines et al. (1979) and Lidicker (1985) reported that juvenile and subordinate males, and lactating females were most likely to disperse. Grid 1 could be near an area with high dispersal and provide a passage for transients because only one juvenile was captured more than once in this grid; the majority of recaptured voles in grid 1 were adults with home ranges within the grid. 17

PAGE 28

3.2 Habitat Characteristics at Capture and Non-capture Sites Ground cover (GCVR), litter depth (LOTH), vegetation height (VEGHGT), and distance from the path (OPTH) were the only variables demonstrating significant differences between capture and non-capture sites in test grids 1 and 2 (Table 3.3). The mean values for GCVR, LOTH, VEGHGT, and OPTH for capture and non-capture sites in all grids are listed in Table 3.4. In grids 1 and 2, voles were captured at sites having denser GCVR, deeper LOTH, taller VEGHGT, and were farther from the path than noncapture locations (Figures 3.1, 3.2). Levels of LOTH and VEGHGT were similar at capture sites in grids 1 (n = 24) and 2 (n = 38) (Wilcoxon ranksum, z = -1.53 (LOTH) and z = -0.34 (VEGHGT), both P values> 0.05). Captures in grids 1 and 2 were associated with 2.44 0.98 em LOTH and 34.92 4.7 em VEGHGT. The average percent GCVR at capture sites in grid 1 was similar to capture sites in grid 2 (Wilcoxon rank-sum, z = 1.86, P > 0.05) with a combined average of 96.54 7.35%. For captures in grid 1, the average distance from the path was 18.10 4.27 m; the average distance from the path for captures in grid 2 was 18.07 5.47 m. No 18

PAGE 29

significant difference in capture site DPTH occurred between grids 1 and 2 (Wilcoxon rank-sum, z = 0.81, P > 0.05). 19

PAGE 30

GCVR LOTH VECVR VEGHGT SPS DALT DSRB DH20 DPTH N Table 3.3 0 GRID1 GRID2 GRID3 CAPTURE VS NONCAPTURE CAPTURE VS NONCAPTURE CAPTURE VS NONCI!PTURE Z value P value Zvalue P value Z value 2.354 0 020* 2 447 0 010* (1. 523) 4.181 <0.002* 2 .741 0.006* (0 184) 0.819 0.420 1 848 0 060 (1.151) 4 376 <0. 002* 2 194 0 030* 0 829 0 .601 0 540 0 706 0 480 0 956 0.079 0.970 1 237 0.210 0 978 1.121 0.260 1.792 0 070 1.178 (0.112) 0 910 0 513 0 610 0.776 2.572 0 010* 2.980 0 002* 0.865 HABITAT CHARACTERISTICS AT CAPTURE AND NON-CAPTURE SITES Comparison of vegetation and landscape characteristics between traps with P value 0.130 0.070 0 250 0 410 0.340 0 330 0.240 0 440 0 380 at least one capture during the study period and traps where no captures were made Differences between capture and non-capture sites were determined by the Wilcoxon rank-sum test and considered significant at P -.5 0.05 Signicant P values are marked with*. Grid 1 captures n=24, non-captures n=36 ; Grid 2 captures n=38, non-captures n=22 ; Grid 3 captures n=53, non-captures n=? Negative values are in parentheses. GCVR = Percent ground cover LOTH= Litter depth (em) VECVR =Percent vegetation cover VEGHGT =Vegetation height (em) SPS = Number of plant species DALT =Shortest distance to different habitat type (m) DSRB =Shortest distance to shrub or tree (m) DH20 =Shortest distance to open water (m) DPTH =Shortest distance to mowed path (m) There is no path in grid 3 ; the DPTH is the distance to the point half way between traps 5 and 6

PAGE 31

N GRID1 GRID2 GRID3 capture noncalm!re calm!re noncalm!re capture GCVR% 98.33 2.93 53.56 32.45 94.74 5.55 86.73 10.4 94.04 7.37 LOTH (em) 2.23 0.99 1.34 0.94 2.67 1.19 1.89 0.87 3.26 1.11 VEGHGT(cm) 34.72 4.51 28.48 5.18 35.03 4.73 30.82 7.41 55.09 15.97 DPTH(m) 18.10.27 12.78 7.60 18.07 5.47 12.27 8.55 15.61 6.92 Table 3.4 MEAN VALUES OF MAJOR HABITAT CHARACTERISTICS AT CAPTURE AND NON-CAPTURE SITES Grid 1 captures n=24, non-captures n=36 Grid 2 captures n=38, non-captures n=22 Grid 3 captures n=53, non-captures n=7 noncalm!re 98.57 2.51 3.34 1.38 51.81 23.78 11.43 5.56 There is no path in grid 3; the DPTH is the distance from a trap to the point half way between traps 5 and 6 in the transect

PAGE 32

Figure 3.1 GRID 1 (TEST) HABITAT CHARACTERISTICS Comparison of statistically significant (P < 0.05) vegetation and habitat features between capture and non-capture sites. Significance determined by Wilcoxon rank-sum test. GRID2 :[ J: 1-a.. 0 CAPS NOCAPS I ffi] GCVR __ ff!j J Figure 3.2 GRID 2 (TEST) HABITAT CHARACTERISTICS Comparison of statistically significant (P < 0.05) vegetation and habitat features between capture and non-capture sites. Significance determined by Wilcoxon rank-sum test. 22

PAGE 33

In control grid 3, no differences were seen between capture (n =53) and non-capture site (n = 7) characteristics (Table 3.3). For all sites in grid 3, the averages were as follows: LOTH 3.27 1.14 em, VEGHGT 55.57 15.55 em, GCVR 94.57 7.12, and DPTH 15.49 6.91% (Figure 3.3). e1oo :r: 80 b ...J 60 0 40 tC) a 20 0 GRID3 ( CAPS NOCAPS [ill GCVR VEGHGT DPTH LOTH Figure 3.3 GRID 3 (CONTROL) HABITAT CHARACTERISTICS Comparison of statistically significant (P < 0.05) vegetation and habitat features between capture and non-capture sites. Significance determined by Wilcoxon rank-sum test. There is no path in grid 3; the DPTH is the distance to the point half way between traps 5 and 6 in the transect. The LOTH and VEGHGT at capture sites in test grids 1 and 2 were significantly less than all sites in control grid 3 (Wilcoxon rank-sum, z =-3.56 (LOTH), z =-7.64 (VEGHGT), both P values< 0.002). No difference occurred among the GCVR at the capture sites in test grids 1 and 2 and all sites in control grid 3 (Wilcoxon rank-sum, z = 0.78, P > 0.05). 23

PAGE 34

At all sites in control grid 3, DPTH was significantly different from the test grids' capture sites (Wilcoxon rank-sum, z = 3.29, P < 0.002). 3.3 Spatial relationship of captures and habitat features Spearman's rank correlation was performed to determine whether the significant variables in grids 1 and 2 (GCVR, LOTH, and VEGHGT-Table 3.3) correlated with the number of captures or distance from the path (Table 3.5). In grid 1, the GCVR, VEGHGT, LOTH, and CAP# increased as distance from the path increased (Figure 3.4). VEGHGT and LOTH were both positively correlated with CAP#; GCVR was not. 100 "#-80 60 40 20 0 GRID1 5 10 15 20 25 DISTANCE FROM PATH ....,.. VEGHGT +"-LOTH -e-GCVR Figure 3.4 GRID 1 (TEST) HABITAT CHARACTERISTICS AND CAPTURE NUMBER CORRELATED WITH DISTANCE FROM THE PATH (DPTH) by Spearman's rank correlation (rs) modified for n > 30 to z values 24

PAGE 35

In grid 2, the GCVR was not correlated with DPTH or CAP# (Figure 3.5). VEGHGT was positively correlated with DPTH, but did not correlate with CAP#. LOTH increased with both DPTH and CAP#. ,----------------------------------------, 100 tl-80 0:: 1j 60 <:> 40 20 0 GRID2 :I: to ...J 0 z L:::::l*": i 5 10 15 20 25 DISTANCE FROM PATH "* CAPt# --...VEGHGT +LOTH --sGCVR ------------------.... -------_____ j Figure 3.5 GRID 2 (TEST) HABITAT CHARACTERISTICS AND CAPTURE NUMBER CORRELATED WITH DISTANCE FROM THE PATH (DPTH) by Spearman's rank correlation (rs) modified for n > 30 to z values In grid 3, no correlation occurred between any of the habitat variables (Figure 3.6). 25

PAGE 36

---------1 100 I .. 20 (!) 0 5 10 15 20 25 DISTANCE FROM PATH Figure 3.6 GRID 3 (CONTROL) HABITAT CHARACTERISTICS AND CAPTURE NUMBER CORRELATED WITH DISTANCE FROM THE PATH (DPTH) by Spearman's rank correlation (rs) modified for n > 30 to z values. There is no path in grid 3; the DPTH is the distance to the point half way between traps 5 and 6 in the transect. 26

PAGE 37

N -....) GRID1 GRID2 GRID3 DPTH CAPt DPTH CAP# DPTH CAP# iltil!l ui.!J!! f.H!.!!!! l...YA!.Yt P yalue GCVR 2.842 0 005* 1.821 0 069 (0. 358) 0.711 1 132 0.256 (0 037) 0.968 (1.460) VEGHGT 3 736 0 .002' 2.470 0 014* 4 210 <0.002* 0 589 0 552 (0.902) 0.358 (0.567) LOTH 3 183 0.002* 3.536 0 002* 3 110 0.003* 3 110 0 003* 0.702 0 472 (1. 553) CAP# ____1.150 0.031* -2 264 0 024* (0 475) -0 631 Table 3 5 CORRELATION OF VEGETATION STRUCTURE. CAPTURE NUMBER, AND DISTANCE FROM THE PATH by Spearman's correlation Correlation coefficients (rs) for sample sizes larger than 30 were converted to two-tailed z values For all distance from path (DPTH) ground cover (GCVR), vegetation height (VEGHGT), and litter depth (LOTH) n=60 Capture number (CAP#) for gid 1(test) n=52 ; grid 2 (test) n=123 ; grid 3 (control) n=227 Negative values are in parentheses Statistically significant values (P < 0 05) are marked with There is no path in the grid 3 (control) ; the DPTH is the distance to the point half way between traps 5 and 6 in the transect 0.144 0.569 0 .121

PAGE 38

3.4 Influence Of The Path On Vegetation Characteristics Characteristics of the trap effective areas incorporating the path originally were calculated including the measurements of the path. The features of the path itself (GCVR 24.01 5.35%, LOTH 0.37 0.15 em, VEGHGT 4.25 1. 75 em) differed in comparison to any other area of the grids (Wilcoxon rank-sum, z = 5.32, 4.32, 4.40 respectively; P < 0.002 for all z values). In order to determine if the portion of the effective area adjacent to the path was influenced by the break in the habitat, the GCVR, LOTH, and VEGHGT were recalculated minus the path data. Within both grids the characteristics of the 4-m strip adjacent to the path were consistent with the poor quality habitat at other non-capture sites greater than 4 m from the path's edge. 3.5 Home Range and Movement Analysis The home ranges of individuals with three or more captures at different trap locations were determined using the inclusive boundary strip method (Stickel 1954). No home ranges in grid 1 (n = 5) or grid 2 (n = 13) included the path or the area less than 4 meters from the path's edge (Figures 3. 7 and 3.8). No trap position effect, relative to the distance between traps 5 and 6, was evident in grid 3. Of the 28 home ranges in grid 3 (Figure 3.9), 22 included some portion of the space between traps 5 and 6 where the 28

PAGE 39

path was located in the test grids. The average home range size in grid 1 was 228.06 32.17 m2 with a maximum distance traveled of 16.80 1.18 m; grid 2 had a 275.51 67.55 m2 range size with a 17.64 3.80 m maximum distance. The grid 3 home ranges averaged 289.42 59.05 m2 with a maximum travel distance of 17.82 2.01 m. The average home range size and maximum distance traveled were compared between grids using the F-test. Grid 1 home range size and distance was significantly smaller than those in grids 2 and 3 (Table 3.6). GRID1 GRID2 HOME RANGE DISTANCE HOME RANGE DISTANCE F value GRID2 8.125 GRID3 10.965 Table 3.6 P value F value P value F value P value F value <0.002* 8.495 0.002* <0.002* 12.468 <0.002* 2.04 0.09 1.468 COMPARISON OF HOME RANGE SIZE AND MAXIMUM DISTANCE TRAVELED by F-test Grid 1 vs grid 2 d. f.: 4/11 Grid 1 vs grid 3d. f.: 4/27 Grid 2 vs grid 3 d.f.: 11/27 Statistically significant P values (P .< 0. 05) are marked with P value 0.193 The number of home ranges correlated with the population density (individuals per hectare) within grids. Of individuals with home ranges, the ratio of males to females was 2:3 in grid 1, 6:7 in grid 2, and 1:1 in grid 3. 29

PAGE 40

Out of 17 4 captures made in grids 1 & 2, no animal switched sides of the path during the study, and only one capture was made in a trap adjacent to the path. The conditions within home ranges were expected to be superior to less frequently traveled sites that might only provide dispersal corridors and short-term resources. To test this hypothesis, the GCVR, LOTH, and VEGHGT within home ranges were compared to variables at captures sites not contained in ranges. Within each grid there were no significant differences in variables between areas within ranges and capture sites not included in ranges. Chi-square analysis and G-test (runs test for randomness) were performed on 12 randomly selected traps and blocks of 12 traps to determine if the capture patterns were significantly different from random events. Each trap location was assigned a number from 1 to 60. A random number generator was used to select traps. Blocks of traps were referenced from a trap selected by the generator. The lack of captures in the 12 traps adjacent to the paths (traps 5 and 6) in grids 1 & 2 was significantly different from what would be expected if capture and non-capture patterns were random events (Chi-square, d.f = 1, P < 0.002; G-test, n = 20, P < 0.002). In contrast, using the Chi-square (d. f. = 1, P < 0.025) and G-tests (n = 20, P < 0.025), only 1 out of 40 30

PAGE 41

capture patterns not including traps 5 and 6 could be differentiated from a random event. All capture patterns tested in grid 3 were not significantly different from random events (Chi-square, d.f = 1, P < 0.002; G-test, n = 20, p < 0.002). 31

PAGE 42

* /I w I ;I N 25 20 Figure 3.7 I )()("' I "' // "' "'"' I )()( p ,, I --r "\ \_j-1')( A T I I H 15 10 5 0 5 10 15 20 25 DISTANCE FROM PATH (meters) GRID 1 (TEST) PRAIRIE VOLE HOME RANGES BY INCLUSIVE BOUNDARY STRIP METHOD Home ranges of females (n=3) delineated with solid lines Home ranges of males (n=2) delineated with dashed lines Each individual is represented by a unique symbol One symbol represents one capturE at that location

PAGE 43

w w JFLL I LLEFFF \ GG \ G JHHHG J GGI-i \ \ GGGJGG GdG \ \ HA 25 20 Figure 3.8 -----1 EE KK BBKK KK / A p \ AA A AA CN -c T \ H / _..:/ > / \ E CM N \ j 15 --1 0 -5 0 5 1 0 15 20 25 DISTANCE FROM PATH (meters) GRID 2 (TEST) PRAIRIE VOLE HOME RANGES BY INCLUSIVE BOUNDARY STRIP METHOD Home ranges of females (n=7) delineated with solid lines Home ranges of males (n=6) delineated with dashed lines Each individual is represented by a unique symbol One symbol represents one capture at that location

PAGE 44

1...1 uut rodd ;pddddo 1 1 s :-;--. / a a ""--' ' yn I I _../ \ ......... I zpp j I l l wphh ., hfcc 25 20 15 10 5 0 5 10 15 20 25 DISTANCE FROM PATH (meters) Figure 3.9 GRID 3 (CONTROL) PRAIRIE VOLE HOME RANGES BY INCLUSIVE BOUNDARY STRIP METHOD Home ranges of females (n=14) delineated with solid lines Home ranges of males (n=14) delineated with dashed lines Each individual is represented by one letter or symbol One letter or symbol represents one capture of that individual at that site Some symbols are moved off-center to improve visibility There is no path in this grid; the presence of a path in the test grids is shaded to allow comparison

PAGE 45

4. Discussion During this study, Prairie Voles did not cross the path, did not include the path in home ranges, and the number of captures increased as the distance from the path increased. It was discovered that a 2-m wide mowed path could induce changes in the edge vegetation structure and inhibit vole dispersion and dispersal as effectively as a road used for vehicular traffic. Prairie Voles in the brome-dominated fields at the Two Ponds National Wildlife Refuge demonstrated a marked preference for habitat providing cover. Individuals were captured, regardless of frequency, in areas with 94.2 5.5% ground cover, average litter depth of 2.50 1.30 em or greater, and standing vegetation cover of 34.19 4.51 em or greater. Habitat with these characteristics is referred to as "high quality" or "preferred" in this paper. Many of the non-capture sites were within the preferred litter depth or vegetation height range but lacked preferred characteristics of both variables. Adequate litter depth and vegetation height appear to work in tandem to contribute to habitat quality. Getz et al. (1978) demonstrated the importance of dense cover to Prairie Voles. They found Prairie Voles thriving along margins of the inhospitable conditions of country roads in Illinois if tall grass was available. Even though mowing was performed on 35

PAGE 46

the road shoulders, the vole populations survived well as long as the activity was temporally and spatially varied. Foster and Gaines (1991), in an investigation of Prairie Vole response to fragmentation, observed aversion to low quality cover, capturing fewer than 1% of animals in a disturbed matrix. High quality habitat is spatially varied through the trapping grids at the Refuge. The contiguous habitat of the control grid (grid 3) was generally high quality; there were no areas larger than 5 by 10 m that could not predictably attract a Prairie Vole. The spatial distribution of captures reflects this homogeneity. The control grid was selected for contiguity of habitat to detect changes in the vegetation structure due to the presence of a path. In grid 3, the average litter depth and vegetation height values exceeded those values in the test grids. However, the discrepancy in vegetation height did not affect the voles' habitat preference. In all grids, the frequency of capture did not correlate with vegetation height after the height reached an average of 34.83 4.54 em. Litter depth increased with the capture number in the test grids, but no correlation was seen between these two variables in the control grid. The threshold for litter depth/capture number correlation occurred between 2.67 1.19 em and 3.26 1.11 em. Within the grids bisected by a path, the habitat quality and capture number increased with distance from the path, but small areas of low 36

PAGE 47

quality habitat were interspersed at farther distances. These variations were probably due to small natural differences in soil chemistry, drainage, or exposure, and would be expected. The only nonrandom dispersion of low quality habitat patches was in the 4-m margin on both sides of the path. The area disturbed by mowing is visually distinct from the border of tall grass on either side but there are subtle differences in the margin habitat that diminish its quality. The abrupt edge created by mowing increases the exposure of the margin to higher temperatures and wind that may result in a shorter, less dense grass crop and low litter accumulation. I made attempts to keep a record of daily temperature highs and lows in the trapping grids. The thermometers were often removed, displaced, broken, or reset by raccoons. I managed to get daily maximum temperatures for clear days in July and August; at the edge of the path daily highs exceeded 50 C; in the grass Sm from the edge the daily highs were 32 to 40 C. This temperature discrepancy may contribute to the variance of vegetation conditions. Based on habitat preference, it was predictable that that no individual included the path and margin in its home range; Swihart and Slade (1984) observed exclusion of a narrow dirt road in Prairie Vole ranges. How does the 10 m of low quality habitat of the path and margins affect dispersal? Other studies have documented the low level of dispersal exhibited by 37

PAGE 48

Prairie Voles in a fragmented habitat even when the suitable patches are small. Diffenderfer et al. (1995) studied the movements of Prairie Voles in an experimentally fragmented field in Kansas. Given a choice of large (50 by 100m), medium (12 by 24m), and small (4 by 8 m) blocks of high quality habitat separated by a matrix ( 16 20 m wide) of low quality mowed habitat, the Prairie Voles selected patch-sizes equally and occupied patches in number proportional to patch size. Switching between patches occurred at a low rate; over an eight-year period greater than 90% of the individuals never switched patches. Once the animals departed the patch, 70% did not return. My study was conducted over a short period (six months), and the effectiveness of the path as a dispersal barrier could not be determined due to a lack of switchers. Population density was proportional to the area of contiguous habitat or patch size. The average maximum range length in grid 1 (16.80 1.18 m) and grids 2 and 3 (17.64 3.80) demonstrate voles' ability to routinely travel a distance necessary to cross the path and margins. In the Diffenderfer et al. (1995) study, the dispersers traveled 1620 m through inhospitable habitat. The cues inducing dispersal are unknown, but a reproductive or density independent population pressure could have an influence. Getz et al. ( 1987) and Gaines and Johnson ( 1984) found that population density is not 38

PAGE 49

an indicator of dispersal pressure. Johnson and Gaines (1987) and Lidicker (1985) reported that dispersing prairie voles tended to be juvenilesubadult males (with high fitness relative to residents), subordinate males, and lactating females. The disproportionate number of males and juveniles captured in grid 1 could be the result of a dispersal effort. The low cover of the path and increased distance between patches with suitable cover increases exposure to predators. Even if individuals were pressured to leave quality habitat, they may not reach a refuge before being eaten. Domestic Cats, Red Foxes, Raccoons, Garter Snakes (Thamnophis radix), and Bullsnakes (Pituophis melanoleucus) have all been seen or have left sign on the trail within the trapping grids. Birds of prey Red-tailed Hawks (Buteo jamaicensis), Swainson's Hawks (Buteo swainsonil), and American Kestrels (Falco sparverius) were all observed hunting in the grass areas. There were not many human visitors to the Refuge, making it unlikely that traffic on the footpaths deterred vole crossings. I was a frequent presence, and groups of 3 to 10 people would briefly tour the area 2 to 3 times a month. If my presence was a factor in the dispersion of the animals, the effect was evenly distributed over the test and control grids. A landscape of small fragmented patches of high-quality habitat is not necessarily a deterrent to the success of Prairie Vole populations. Prairie Voles can adapt and utilize small areas to their advantage. Diffenderfer et 39

PAGE 50

al. (1995) reported an advantage by female Prairie Voles in defending home ranges surrounded by low quality matrix. The Prairie Vole, in contrast to the Meadow Vole is a social animal mates form pair bonds, older offspring care for neonates, and range overlap between non-related groups may occur (Getz 1962; Thomas and Birney 1979; Getz and Pizzuto 1987). Social dynamics may change with the population number. Territoriality and aggression may occur between adults of both sexes in reproductive condition, or as population densities increase, territoriality may be relaxed and a promiscuous or polygynous system ensues (Rose and Gaines 1976; Getz et al. 1987). The Prairie Vole populations at Two Ponds appear to be operating at or below carrying capacity because the sex ratios of individuals with home ranges within grids is similar (a slight skew to the female side noted in grids 1 and 2) strongly suggests monogamy and territoriality; the home range size of individuals within grids is similar despite the large discrepancy in population density and biomass. Prairie Vole home range size was negatively correlated with population density and habitat biomass in studies performed by Abramsky and Tracy (1979). If the carrying capacity were exceeded in an area, a breakdown of monogamy, territoriality, or a decrease in home range size would be evident. 40

PAGE 51

The trail system in the grassland could have long-term effects on Prairie Vole populations by decreasing available habitat and increasing distance between suitable habitat patches. Diffenderfer et al. (1995) reported that voles' dispersal rates decreased with increasing distance between suitable habitat patches, resulting in isolation of populations. Isolation is associated with an increased incidence of inbreeding with subsequent loss of fitness and genetic diversity (Ralls et al. 1986; Barrett and Kohn 1991 ). The concentration of populations in small isolated areas increases their susceptibility to disease and catastrophic events (e.g. floods, fire, mowing). As distance between suitable patches increases, the chance of areas being recolonized by the same species diminishes. Two Ponds NWR is isolated by development; the most likely colonizer of disturbed habitat would be generalists like House and Deer Mice. 41

PAGE 52

5. Conclusion Mowed paths do not provide suitable Prairie Vole habitat and jeopardize the quality of the adjacent areas due to edge effect. Although the vegetation structure of the 4-m wide margins appears distinctly different from the 2-m wide path, increased exposure has resulted in diminished litter depth and grass height. The 1O-m wide strip of path and associated low quality margin reduces available habitat in the Refuge's grassland by approximately 1. 0 ha per km of path or 21% of the total area. The increase in habitat fragmentation and distance between suitable habitat patches raise the cost of dispersal, although Prairie Voles tolerate small habitat patches, have low dispersal rates, and avoid areas that lack cover. The path and associated margins pose a threat to M. ochrogaster by facilitating predator access, reducing available habitat, and contributing to a reduction of carrying capacity and decline of population size. 42

PAGE 53

6. Management Recommendations Maintaining healthy Prairie Vole populations at Two Ponds National Wildlife Refuge should be encouraged. This native vole is an important prey species for raptors, carnivorous mammals, snakes, and amphibians. Prairie Voles can inhibit the invasion of non-native species, such as House Mice, and a hantavirus carrier-the Deer Mouse (Peromyscus maniculatus) (Abramsky and Tracy 1979). If the primary management goal were to maintain wildlife habitat, the abandonment of the current trail system is suggested. The Refuge is a venue for education classes; therefore, access is required which minimizes impact on the environment. The influence of the trail system on the grassland can be lessened as follows: 1) Establish paths that end in a cul-de-sac and do not circle and isolate large patches. 2) Install bridges or boardwalks over the path (up to1 0 em high) to provide providing cover. Ten centimeters is recommended because I observed vole tunnels crossing paths that were covered with 1 0 em of snow, and access to large predators is discouraged. There are already four bridges of adequate 43

PAGE 54

dimensions (3-m long, 2-m wide) in the west area, but the drainage ditches under the structures are used as passageways by raccoons and foxes. Monitoring the effect of bridge or boardwalk structures on vole dispersion and vegetation structure is an idea for a future field study. 3) Irrigate the path's 4-m margins to diminish the dessicating edge effect. Abdellatif et al. ( 1982) saw increased biomass and vole reproductive success when supplementary water was applied to a grassland. 4) If paths are paved, use a material that minimizes heat accumulation; therefore, diminishing the effect of increased temperatures. 5) The effect of decreasing the width of the path is a subject for further study. Decreasing the path's width to 1 m will allow pedestrian access, decrease the distance between suitable patches, and possibly alleviate the edge effect. The impact of planned habitat disturbancesconstruction, weed control, or vegetation restoration alterations should be carefully evaluated and mitigation performed if possible. Grassland alterations should be spatially and temporally varied to provide sanctuary for displaced animals. Habitat should be provided with 94.20 5.50% ground cover, litter depth greater 44

PAGE 55

than 2.50 1.30 em, and a growing season vegetation height of 34.72 4.51 em or greater. In this study, shrubs (rabbitbrush, Chrysothamnus nauseosus) in the grassland, did not contribute to the presence or increased number of Prairie Voles. Periodic surveys for species diversity, population densities, and distribution should be performed to assess the impact of anthropogenic disturbances and efficacy of management techniques for maintaining wildlife habitat at Two Ponds National Wildlife Refuge. 45

PAGE 56

References Abdellatif, E.M., K.B. Armitage, M.S. Gaines, and M.L. Johnson. The effect of watering on a prairie vole population. Acta Theriologica 27: 243255. Abramsky, Z., and C.R. Tracy. 1979. Population biology of a "noncycling" population of prairie voles and a hypothesis on the role of migration in regulating microtine cycles. Ecology 60: 349-361. Armstrong, D.M. 1972. Distribution of mammals in Colorado. University of Kansas, Museum of Natural History Monographs 3: 1-415. Barrett, S.C. H., and J.R. Kohn. 1991. Genetic and evolutionary consequences of small population size in plants: Implications for conservation. Pages 3-30 in D.A. Falk and K.E. Holsinger, editors. Genetics and Conservation of Rare Plants. Oxford University Press, New York. Bierregaard, R.O., T.E. Lovejoy, V. Kapos, A.A. DosSantos, and R.W. Hutchings. 1992. The biological dynamics of tropical rainforest fragments. BioScience 42: 859-866. Burt, W.H. 1943. Territoriality and home range concepts as applied to mammals. Journal of Mammalogy 24: 346-352. Carey, A B. 1978. Distributional records for the prairie vole and hispid cotton rat in Colorado. Journal of Mammalogy 59: 624. Carroll, D., and L.L. Getz. 1976. Runway use and population density of Microtus ochrogaster. Journal of Mammalogy 57: 772-776. Carver, A.V., L.W. Burger, and L.A. Brennan. 1999. Passive integrated transponders and patagial tag markers for Northern Bobwhite chicks. Journal of Wildlife Management 63(1): 162-166. 46

PAGE 57

Collins, R.J., and G.W. Barrett. 1997. Effects of habitat fragmentation on meadow vole (Microtus pennsylvanicus) population dynamics in experimental landscape patches. Landscape Ecology 12(2): 63-76. Curtis, J.T. 1956. A prairie continuum in Wisconsin. Ecology 36: 558-566. Diffenderfer, J.E., M.S. Gaines, RD. Holt. 1995. Habitat fragmentation and movements of three small mammals ( Sigmodon, Microtus, and Peromyscus). Ecology 76(3): 827-839. Dooley, J.L., and M.A. Bowers. 1998. Demographic responses to habitat fragmentation: Experimental tests at the landscape and patch scale. Ecology 79(3): 969-980. Elbin, S.B., and J. Burger. 1994. Implantable microchips for individuals identification in wild and captive populations. Wildlife Society Bulletin 22: 677-683. Fagerstone, K.A., and B.E. Johns. 1987. Transponders as permanent identification markers for domestic ferrets, black-footed ferrets, and other wildlife. Journal of Wildlife Management 51: 294-297. Findley, J.S. 1954. Competition as a possible limiting factor in the distribution of Microtus. Ecology 35: 418-420. Fisher, H.J. 1945. Notes on voles in central Missouri. Journal of Mammalogy 26: 435-437. Forman, R. T. T., and M. God ron. 1986. Landscape Ecology. John Wiley and Sons, New York. Forman, R.T.T., and RD. Deblinger. 2000. The ecological road-effect zone of a Massachusetts (U.S.A.) suburban highway. Conservation Biology 14(1 ): 36-46. Foster, J., and M.S. Gaines. 1991. The effects of a successional habitat mosaic on a small mammal community. Ecology 72: 1358-1373. 47

PAGE 58

Gaines, M.S., A.M. Vivas, and C.L. Baker. 1979. An experimental analysis of dispersal in fluctuating prairie vole populations: demographic parameters. Ecology 60: 814-828. Gaines, M.S., and M.L. Johnson. 1984. A multivariate study of the relationship between dispersal and demography in population of Microtus ochrogaster in eastern Kansas. The American Midland Naturalist 111: 223-233. Getz, L. L. 1962. Aggressive behavior of the meadow and prairie vole. Journal of Mammalogy 43: 351-358. Getz, L.L., F.R. Cole, and D.L. Gates. 1978. Interstate roadsides as dispersal routes for Microtus pennsylvanicus. Journal of Mammalogy 59(1 ): 208-212. Getz, L.L., J.E. Hofmann, and C.S. Carter. 1987. Mating systems and population fluctuations of the prairie vole, Microtus ochrogaster. American Zoologist 27: 909-920. Getz, L.L., and T.M. Pizzuto. 1987. Mating system, mate preference and rarity of blond prairie voles. Transactions of the Illinois Academy of Sciences 80: 227-232. Hamner, B. 1989. PIT tags for animal husbandry and visitor education systems. Proceedings of the American Association of Zoological Parks and Aquariums Reg. Meetings, Wheeling, West Virginia: 443447. Haskell, D.G. 2000. Effects of forest roads on macroinvertebrate soil fauna of the southern Appalachian Mountains. Conservation Biology 14( 1 ): 57-63. Hayne, D. W. 1949. Calculation of home range size. Journal of Mammalogy 30(1): 1-16. Hayne, D.W. 1950. Apparent home range of Microtus in relation to distance between traps. Journal of Mammalogy 51 (1 ): 26-39. 48

PAGE 59

Johnson, M.L., and M.S. Gaines. 1987. The selective basis for dispersal of the prairie vole, Microtus ochrogaster. Ecology 68:684-694. Kapos, V. 1989. Effects of isolation on the water status of forest patches in the Brazilian Amazon. Journal of Tropical Ecology 5: 173-185. Kasworm, W., and T. Manley. 1990. Road and trail influences on grizzly and black bears in northwest Montana. International Conference on Bear Research and Management 8: 79-84. Korschgen, L.J., and T.S. Baskett. 1963. Food of impoundmentand stream dwelling bullfrogs in Missouri. Herpetologica 19: 89-99. Laurance, W. F. 1991. Edge effects in tropical forest fragments: Application of a model of nature reserves. Biological Conservation 57: 205-219. Lidicker, W.Z., Jr. 1985. Dispersal. Pages 420-454 in R.H. Tamarin, editor. Biology of New World Microtus. Special Publication of The American Society of Mammalogists 8: 1-893. Lynch, J.F., and D.F. Whigham. 1984. Effects of forest fragmentation on breeding bird communities in Maryland, USA. Biological Conservation 28: 287-324. Mader, H.J. 1984. Animal habitat isolation by roads and agricultural fields. Biological Conservation 29: 81-96. Martin, E.P. 1956. A population study of the prairie vole (Microtus ochrogaster) in northeastern Kansas. Miscellaneous Publication of the Museum of Natural History, University of Kansas 8: 361-416. Moulton, M.P., J.R. Choate, S.J. Sissel, and R.A. Nicholson. 1981. Associations of small mammals on the central high plains of eastern Colorado. Southwestern Naturalist 26: 53-57. Nassauer, J .I. 1995. Culture and changing landscape structure. Landscape Ecology 10: 229-237. 49

PAGE 60

Nupp, T. E., and R. K. Swihart. 1996. Effect of forest patch area on population attributes of white-footed mice (Peromyscus leucopus) in fragmented landscapes. Canadian Journal of Zoology 74: 467-472. Owensby, C.E. 1973. Modified step-point system for botanical composition and basal cover estimates. Journal of Range Management 26: 30.2303. Palmer, E.L., and H.S. Fowler. 1975. Fieldbook of Natural History, second edition. Page 712. McGraw-Hill Book Company, New York Ralls, K., P.H. Harvey, and A.M. Lyles. 1986. Inbreeding in natural populations of birds and mammals. Pages 35-56 in M. E. Soule, editor. Conservation Biology: The Science of Scarcity and Diversity. Sinauer Associates, Sunderland, Maryland. Ribble, D.O. and F.B. Samson. 1987. Microhabitat associations of small mammals in southeastern Colorado, with special emphasis on Peromyscus (Rodentia). Southwestern Naturalist 32(3): 291-303. Reed, M.P., and J.R. Choate. 1988. Noteworthy southwestern records of the prairie vole. Southwestern Naturalist 33: 495-496. Rose, R. K., and M.S. Gaines. 1976. Levels of aggression in fluctuating populations of Microtus ochrogaster, in eastern Kansas. Journal of Mammalogy 57: 43-57. Saunders, D.A., R.J. Hobbs, and G.W. Arnold. 1993. The Kellerberrin project on fragmented landscapes: a review of current information. Biological Conservation 64: 185-192. Shriner, W.M., and P.B. Stacey. 1991. Spatial relationships and dispersal patterns in the rock squirrel, Spermophilus variegatus. Journal of Mammalogy 72(3): 601-606. 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. 50

PAGE 61

Slade, N.A, and L. A Russell. 1998. Distances as indices to movements and home-range size from trapping records of small mammals. Journal of Mamma logy 79( 1 ): 346-351. Stalling, D.T. 1990. Microtus ochrogaster. Mammalian Species 355: 1-9. Stalling, D.T. 1999. Pages 637-638 in D.E. Wilson and S. Ruff, editors. The Smithsonian Book of North American Mammals. Smithsonian Institution Press, Washington, D.C. Stickel, L.F. 1954. A comparison of certain methods of measuring ranges of small mammals. Journal of Mammalogy 35(1 ): 1-15. Swihart, R.K., and N.A Slade. 1984. Road crossing by Sigmodon hispidus and Microtus ochrogaster. Journal of Mammalogy 65: 357-360. Szacki, J., J. Babinska-Weka, and A Lira. 1993. The influence of landscape spatial structure on small mammal movements. Acta Theriologica 38:113-123. Thomas, J.A, and E.C. Birney. 1979. Parental care and mating system of the prairie vole, Microtus ochrogaster. Behavioral Ecology and Sociobiology 5: 171-186. Thomas, J.W., C. Maser, and J.E. Rodick. 1979. Edges. Pages 48-59 in J.W. Thomas, editor. Wildlife Habitats in managed forest: The Blue Mountains of Oregon and Washington. U.S. Department of Agriculture, Forest Service, Agriculture Handbook No. 553, Washington, D.C. Villard, M., M.K. Trzcinski, and G. Merriam. 1999. Fragmentation effects on forest birds: relative influence of woodland cover and configuration on landscape occupancy. Conservation Biology 13(4): 774-783. Webster, W.O., and J.K. Jones. 1982. Reithrodontomys mega/otis. Mammalian Species 167: 1-5. Wilcove, D.S. 1985. Nest predation in forest tracts and the decline of migratory songbirds. Ecology 66(4): 1211-1214. 51

PAGE 62

Yahner, R.H. 1988. Changes in wildlife communities near edges. Conservation Biology 2: 333-339. 52