NEOTOMA FLORIDANA FLORIDANA NATURAL HISTORY, POPULATIONS, AND MOVEMENTS IN NORTH-CENTRAL FLORIDA Leslie HaySmithi ABSTRACT A behavioral ecology study on the eastern woodrat (Neotoma floridana foridana) was conducted between September 1988 and February 1990 on the Katharine Ordway Preserve-Swisher Memorial Sanctuary. The Preserve is located in north-central Florida, approximately 5 miles east of Melrose. The goal of the study was to investigate factors that may influence density and distribution of Neotoma on the Ordway Preserve. Distribution of the eastern woodrat is closely associated with mesic hardwood hammocks on the Preserve. Five study sites were therefore selected in three habitat types: mesic hammock with open understory, mesic hammock with saw-palmetto understory, and bottomland hardwood swamp. Trapping along strip-transects at these sites demonstrated that the bottomland hardwood swamp consistently had the highest density of woodrats (4.3/ha at initiation of study). Densities fluctuated in all 5 study sites, and at Goose and Ashley lakes populations crashed. Radio telemetry was used to evaluate woodrat movements, habitat utilization, and den use. Woodrats were nocturnal with peaks in nighttime activity between 2000-2200 and 0100-0330 hours. Mean home range size was larger in the mesic hammock with saw-palmetto understory (1.05 ha), which was least dense in horizontal woody growth. Home ranges were smaller in the mesic hammock with open understory (0.25 ha) and bottomland hardwood swamp (0.69 ha) which were more complex in horizontal and vertical vegetation structure. Home range size was positively correlated to number of dens used, which ranged from 1 to 3 dens in use at one time. Dens of woodrats were the center of all activity. Nest fidelity was high, but some individuals changed dens. Social tolerance was low, radiotelemetry and trapping observations rarely documented animals in close proximity to each other. In summary, the bottomland hardwood swamp, which has the highest stein density and vegetation cover supports the highest density ofwoodrats. These habitat features may influence smaller home ranges. The density fluctuations in all study sites and subsequent population crashes in two mesic hammock habitats, suggest woodrats may be influenced by external factors such as predation and drought conditions. Forests with less horizontal and vertical structural complexity may be marginal to woodrats due to cover requirements and food resource distribution. Populations inhabiting these environments may therefore be more vulnerable when heightened stress levels occur, such as changes in climatic conditions. 1 The author is the Coordinator of International Programs in the College of Forestry, Wildlife and Range Sciences, University of idaho. Mosco ID 83843, U SA HAYSMITH, L. 1995. Neotoma/loridana~#oridana: Natural history, populations, and movements in north-central Florida. Bull. Florida Mus. Nat. }list 38 PL Il(8):211-243. 212 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL- 38 PT. II(8) RESUMEN Se realiz6 un estudio sobre la ecologia conductual de la mla de bosque del este (Neotomaf?oridana flondana) entre septiembre de 1988 y febrero de 1990, en la Reserva Katharine Ordway-Swischer Memorial Sanctuary. La Reserva esti localizada en el norte-centro de Florida aproximadamente 5 millas al este de Melrose. El objetivo de este estudio fue investigar los factores que pueden influenciar la densidad y distribuci6n de Neotoma en la Reserva Ordway. La distribucidn de la rata de bosque del este en la Reserva esth fuertemente asociada con los bosques demadera duram6sicos. Por lotanto, se seleccionaron cinco sitios de estudio en tres tipos dehabitat: bosque m6sico con sotobosque abierto, bosque m6sico con sotobosque de palma y pantano bajo de madera dura. El trampeo a 10 largo de transectos en estos sitios, demostr6 que los pantanos bajos de madera dura tuvieron consistentemente la densidad mds alta de ratas de bosque (4.3/ ha en la iniciaci6n del estudio). La densidades poblacionales fluctuaron en los cinco sitios de estudio, colapsando 6stas en los lagos Goose y Ashley. Se us6 radiotelemetria para evaluar los movimientos, el uso de habitat y de madrigueras de las ratas de bosque. Las ratas de bosque fileron nocturnas, con mixima actividad nocturna entre 2000-2200 y 0100- 0330 horas. La media del tamailo del rango de hogar fue mayor en los bosques m6sicos con sotobosque de palmas (1.05 ha), habitat que fue el menos denso en crecimiento leaoso horizontal. Los ambitos de hogar fueron mds pequeaos en el bosque m6sico con sotobosque abierto (0.25 ha) y en el pantano bajo de madera dura (0.69 ha), teniendo ambos una estructura vegetal mAs compleja horizontal y verticalmente. El tamaito del dmbito de hogar estuvo positivamente correlacionado con el ndmero de madrigueras utilizado, variando ask de entre la3 madrigueras utilizadas cada vez. Las madrigueras de las ratas de bos(lue fueron el centro de tola activiclad La fidelidad a 6stas fue alta, aunque algunos individuos cambiaron madrigueras. La tolerancia social fue baja; datos de telemetria y trampeo ran vez documentaron animales estrechamente cercanos unos de otros. En resumen, el panlano bajo de madera dura, el cual tiene la mayor densidad de ramas y cobertura vegetacional, mantiene la mayor densidad de ratas de bosque. Estas caracteristicas del habitat probablemente influencian dmbitos de hogar mhs pequeflos. L.as fluctuaciones poblacionales en todos los sitios de estudio, y el subsecuente colapso en dos habitats de bosque m6sico, sugieren que las ratas de bos(lue pue(len estar influenciadas por predaci6n y condiciones de sequia en estos habitats. L,os bosques con menor compleji(laci estructural horizontal y vertical pueden ser marginales para las ratas de bosque debido a sus requerimientos de cobertura y distribuci6n de recursos alimenticios. Las poblaciones que habitan estos ambientes pueden, por 10 tanto, ser mas vulnerables a mayores niveles de estrds, tales como cambios en las condiciones climMicas. INTRODUCTION Few comprehensive studies on the ecology of the Eastern woodrat (Neotoma floridana) have occurred, despite its widespread geographical distribution in the eastern United States. Nine subspecies are currently recognized in this geographical range (Wiley 1980). Neotoma f floridana is the southeasternmost subspecies located in Florida, Georgia, South Carolina, and North Carolina. Woodrats have a long life expectancy and low reproductive rate compared to other small rodents (Hamilton and Whitaker 1981; Hamilton 1953). They are omnivorous, consuming a variety of invertebrates and plants (Hamilton and Whitaker 1981). Dens are a significant component in Neotoma floridana ecology. Movements are closely linked to their den, which is the base of all their operations (Fitch and Rainey 1956). Controversy exists on whether the woodrat is solitary or colonial. Some researchers indicate Neotoma floridana are colonial, communally occupying dens OPearson 1952; Goertz 1970), and others list the species as strictly solitary and territorial (Poole 1940; Hamilton and Whitaker 1981). HAYSMITH: NEOTOMA F. PLORIDANA IN NORTH-CENTRAL FLORIDA 213 WHe Neotoma jtoridana has a broad geographical range, it appears to have a patchy distribution throughout the range for unknown reasons (Schwartz and Odum 1957). Habitat requirements of this species are poorly understood; it is found in rocky cliffs, ravines, upland oak forests, hardwood bottomland forests, riparian woods, and lowland wet hammocks (Hamilton and Whitaker 1981). Their presence in widely different habitats throughout the eastern United States creates difficulty in understanding if optimal or suboptimal habitats can be distinguished and what specific habitat features are necessary to meet life history requirements. In peninsular Florida mammal distributions are very complex (Eisenberg 1989). Neotoma floridana distribution is likewise unclear, but it appears to occur throughout most of peninsular Florida. Layne (1974) indicated the range did not extend south of Lake Okeechobee, and Greer (1978) discovered woodrats in highly localized populations south to DeSoto County. Greer indicated south Florida Neotoma populations are disjunct for unknown reasons and speculated that coastal development may be influencing their distribution. Populations of the eastern woodrat occur in a wide range of densities throughout its geographical distribution in the United States. Numerous researchers have attributed different densities to habitat conditions, food availability, predation pressure, and other ecological factors (Worth 1950; Pearson 1952; Goertz 1970). However, it is not clear what factors may be most important in influencing the highly variable densities. While densities vary throughout its range, populations of the eastern woodrat appear to be stable in some areas, and vulnerable in others. Woodrats are endangered in Key Largo, Florida (subspecies, N. f smalli; Humphrey 1988) and threatened in New Jersey, eastern Pennsylvania, and New York, where it has completely disappeared from much of its historical range (Sciascia 1990). While the woodrat is not threatened in most of the southeast, native habitats of mesic hardwood forests in Florida which appear to be the principal habitat are rapidly disappearing. Studies are needed to understand life history parameters, and those that influence their density and distribution. Incongruent reports on the life history of Neotoma floridana prompted this investigation into population parameters, movement patterns, and habitat utilization on macro and micro levels. My goal was to investigate parameters that influence the density and distribution of Neotoma f foridana in mesic forest habitats on the Ordway Preserve. I had four chief objectives in this study: (a) to assess woodrat population dynamics among and within mesic forests at the Ordway Preserve; (b) to identify whether this sub-species exhibits habitat preferences among and within mesic forest types; (c) to determine home range size, pattern, and nest use; and (d) to assess whether movement and nesting activity may be related to specific habitat features. 214 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL, 38 PT. II(8) ACKNOWLEDGEMENTS I would like to thank Margie Voeten, Leslie Straub, and Kathy Hallman for field assistance. John F. Eisenberg Richard Franz, and Melvin E. Sunquist provided valuable advice on study design and analysis. I express gratitude to the Ordway Foundation for logistical and financial support for the duration of the project. I would also like to thank the reviewers for their time and effort in reviewing this manuscript. STUDY AREA The Katharine Ordway Preserve/Swisher Memorial Sanctuary was established as a reserve in 1979. Because of its brief history of protection mixed historical use has resulted in a composite landscape of contiguous expanses of native xeric and mesic forests, wetlands, and prairies, as well as fragmented forest patches, bounded by old fields and former pastures. The habitat fragmentation, particularly evident near black water lakes, is a result of agricultural, pastoral, and silvicultural activities from the last century. The complexity of topography and soil conditions, the position of the Ordway Preserve on the Central Florida ridge, and past human perturbations lead to a unique assemblage of vascular plant flora (Franz and Hall 1990). Major communities represented are: high pine forests, sand live oak hammocks, mesic hardwood hammocks, swamp forests, freshwater marshes, sandhill lake fringes, permanent lakes and ponds, and culturally derived sites (Franz and Hall 1990). METHODS This study was conducted over a 16-month study period from September 1988 to January 1990. I selected five study sites in three habitat types: the bottomland hardwood swamp (Mill Creek Swamp), mesic hammock with saw-palmetto understory (Ross and Goose lakes), and mesic hardwood hammock with open understory (Ashley and Suggs lakes) (Fig. 1). The bottomland swamp is a large continuous tract of forest with numerous tree species and dense ericaceous shrubs. The mesic hammock with saw-palmetto understory (Goose and Ross lakes) and HAYSMITH: NEOTOMA F. FLOR/DANA IN NORTH-CENTRAL FLORIDA 215 ORDWAY PRESERVE 11%3-3~ N ; %5 0 1000 meters i ~0 , uir·,4 A 71, 9\1_ M.r r~ 6 1 0 . )*3/ *A. 43 2 'YA t...4, 05 61 ~1.< __3* 1 = Ashley Lake 2 = Goose Lake 3 = Ross Lake MH = mesic hardwood hammock 4 = Mill Creek Swamp MHP = mesic hammock saw palmetto 5 = Suggs Lake BSW = bottomland hardwood swamp • = Distribution of Neoroma on the Ordway Black = mesic hardwood forests Figure 1. Mesic forests, study areas, and distribution ofNeotomafforidana on the Ordway Preserve. mesic hardwood forests with open understory (Suggs and Ashley lakes) are narrow belts of mesic forest on the perimeter of the black water lakes which are 50-200 m in width and are bounded by the lake, a road, and old fields, resulting in fragmented habitats. Throughout the last decade other mammalian surveys (over 10,000 trap nights) have been conducted in various habitats on the Preserve. The results of these surveys indicated that woodrats were primarily distributed in mesic forests. I selected my five study sites based on previous records of woodrats at these sites. Vegetation measurements were collected using the point-center-quarter method (Hays et al. 1981; Ludwig and Reynolds 1988). Because of the linear nature of the habitats, two randomly located transects were established at 90° angles in each study site, which were parallel and perpendicular to the lakes. A vegetation station, which was divided into four quarters, was established every 50 m. 9 Woodrats were captured in squirrel size Tomahawk traps (61 x 17 x 17 cm) baited with a mixture of peanut butter and oatmeal. Standard biological measurements were taken which included: weight, head and body length, tail 216 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 38 PT. II(8) length, and right hind foot length. Data were also collected on gender, age class (adult or juvenile), and reproductive condition. Ectoparasites were evaluated and collected. Animals were initially marked with ear tags (#lL, National Band and Tag Company, Newport, KY), but were later marked with ear notches using a standard numbering system, due to problems with infections. Traps were placed at 15-20 m intervals along 400 m strip transects near vegetation, logs, and in brushpiles. Transects were established parallel to lake edges due to the linear nature of the study sites. The two parallel strip transects were set on random compass bearings, intersecting microhabitats and ecotonal areas. Trapping effort was systematically varied throughout the 16-month study period, corresponding to twice per month in each study site. Trapping schedule was 4-5 days per session. In addition to trapping on transects, den sites were trapped. Ecological densities, defined as the number of individuals per unit habitat, were calculated as the minimum number of animals known alive (MNKA) per hectare (capture of the total population could never be assumed). Density was calculated as the number of individuals captured per unit area. Total area was calculated using aerial photos. I also evaluated 42 woodrat dens for 21 structural and vegetative characteristics in all three habitat types. At each den measurements were taken on: location of den, substrate species, height, DBH (diameter basal height), cover, distance to water, and debris at den. Selected individuals were equipped with 10 gram radio collars. Transmitters and collars consisted of a small mercury oxide hearing aid battery, an electrical board, and a rubberized collar, which was covered with epoxy and electrical tape. The total weight of the transmitter and collar was less than 5% of the total body weight. A total of 27 animals (16 males and 11 females) were radiotracked. Individual animals were tracked for periods varying from two weeks to six months. Telemetry data were collected from 1800 to 0600 EST. To obtain data on activity periods, a reading was taken on each animal every two hours. Animals were radiotracked by triangulation and the "homing in" technique (Harris et al. 1990). The latter method was primarily used in daylight hours when animals were in their dens. Data recorded using triangulation were: date, individual animal and frequency; time, station number, and location; azimuth; and animal activity. The estimated degree of error with triangulation was 7.25 m, forming a relatively small error radius around the estimated location of the animal. Home range data were analyzed using the minimum convex polygon (MCP) and modified concave polygon methods (COC). The MCP is the 'total area of activity' following Burt (1943). The COC is the core area, or "central area of consistent or intense use" (Kaufmann 1962) whereby 50% of the maximum distance between points is utilized (Michener 1979). The concept of center of activity in home range analysis (Harris, et al. 1990) is simplified with woodrats, HAYSMITH: NEOTOA£.4 E FLORIDANA IN NORTH-CENTRAL FLORIDA 217 because woodrat dens are the center of all their activity, as documented in this and other studies. Home ranges were analyzed by individual animal, gender, daily and monthly time periods, and habitat type. The minimum convex polygon is the traditional method used in radiotelemetry studies. It is constructed by connecting the outermost animal locations, and calculating the area within this polygon. The modified concave polygon is similar to the minimum convex polygon, but uses a prespecified maximum distance between perimeter points in the analysis, eliminating outlying points (Harvey and Bahour 1965). Observations that are farthest from the arithmetic center are thus removed and several use areas are produced, therefore the modified concave polygon (COC) method yields the smallest ranges. Fluorescent pigments were used to evaluate three-dimensional movements and habitat utilization. Two methods were implemented: full-body dusting and attachment of fluorescent capsules. Full-body dusting was done by placing an animal in a bag with fluorescent powder, lightly shaking the bag with the woodrat, and immediately releasing the animal. The capsule method involved fabricating a small capsule from paraffin wax, and dental acrylic (Goodyear 1989), injecting powder, and gluing the capsule to the animals pelage. Powder trails were followed with a battery operated UV light, using stakes and flags to mark the trails (Leman and Freeman 1985; Goodyear 1989). Micro-habitats in heterogenous mesic forests were mapped in the field to obtain information on understory strata not discernible on aerial photographs. These maps were later transformed to habitat maps drawn to the output scale of the Telem88 home range program (Coleman and Jones 1988) for plotting animal home ranges, with delineations set at proper compass bearings. Utilization of habitat by woodrats was assessed by plotting all animal home range locations and overlaying these points onto the habitat maps, using X and Y coordinates. The radiotelemetry locations were thereby assigned to a habitat, and points were tallied for each of these categories (Sunquist 1989). RESULTS Habitat Ground, herbaceous, and canopy cover, in addition to litter depth was evaluated between the three habitat types. Parametric tests (ANOVA) and non- parametric tests (Kruskal-Wallis) were used to determine if differences existed between habitat types for vegetative parameters. Mean values for the three types of cover and litter depth were similar, and an ANOVA demonstrated no significant differences for cover and litter between habitats *=0.05). Species composition of plants between the three habitat types were similar in canopy, midstory, shrub, and herbaceous strata. Tree species diversity was high in 218 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 38 PT. II(8) all three habitats, with each habitat including between 9 and 11 species. Dominant tree species in the canopy, midstory and ground cover strata were water oak (Quercus nigra), live oak (Quercus virginiana, laurel oak (Quercus laurifolia, sweet gum U.iquidambar styracijlua), swamp bay (Persea palustris), sweet bay (Magnolia virginiana), and pignut hickory (Caga glabra) (Tables 1, 2 and 3). The shrub layer differed the most in species composition among the three habitat types. Dominant shrubs were saw-palmetto (Serenoa repens), fetterbush (Lyonia /ucida), and wax myrtle (Myrica cer(tera). The most prevalent difference occurred in the bottomland swamp (Mill Creek Swamp) with the predominance of ericaceous shrubs (Table 4). Herb composition within the habitats was more diverse than other strata. Several species of ferns were present in all three habitats. Other species of herbs and grasses occurred in highest abundance in the mesic hardwood hammock with open understory (Table 5). Several species of vines, including Smilax and muscadine grape, occurred in nearly all sites (Table 6). Relative densities of plant species in all three habitat types in the three strata were similar. Relative density of trees, saplings, and seedlings were concentrated in four species with a range between 12% and 31% (Tables 1, 2, and 3). Differences in relative densities for shrubs were most pronounced in the mesic hammock with saw-palmetto understory (Ross and Goose lakes) and bottomland hardwood swamp habitats with palmetto (44.4%) and fetterbush (54.2%) (Table 4). Concentrations of relative densities for herbs and vines occurred in four species, ranging between 13% and 38% (Tables 5 and 6). The most significant difference between habitat types was vertical and horizontal structure, rather than species composition or relative density. Structural complexity of vegetation in horizontal and vertical planes measured by height, stem density, and crown diameter varied considerably among habitat types. The mean values of vegetation height in the three habitat types are presented in Table 7. An ANOVA of vegetation height demonstrated significant differences in all strata for saplings and shrubs by habitat where saplings were tallest in the bottomland swamp and shortest in mesic hardwoods; shrubs were tallest in mesic hammock with saw-palmetto understory. Trees, seedlings, and herbs showed no significant difference in height by habitat type (p=0.05). Diameter basal height (DBH) was measured as an index of plant dominance. An ANOVA demonstrated significant differences for DBH of trees, saplings, and shrubs between habitats (p=0.05). Mean DBH for trees was highest in the mesic hardwood forests; shrub DBH was highest in mesic hammock-saw-palmetto; and the highest DBH for saplings occurred in the bottomland swamp (Table 8). HAYSMITH: NEOTOMA F. FLORIDANA IN NORTH-CENTRAL FLORIDA 219 Table 1. Relative density (%) oftrees in three habitat t>~es on the Onlivay Preserve Sites' MH MHP BSW Habitat Species SUG ASH ROS GOS MCS Uve oak 20.00 14.29 18.18 12.77 (Quercus vi*mana) Swed gum 30.95 12.50 10.64 (Jiquidambar styracijlua) Pignut hickory 19.05 (Carya glabra) Water oak 16.67 6.66 21.43 38.63 21.28 Quercus nigra) Black gum 7.14 1.79 13.63 (Nyssa sylvalica) Longleafpine 2.38 2.27 21.28 (Pinus palustris) Swampbay 2.38 20.00 26.79 8.51 (Persea palustris) Dahoon holly 2.38 3.57 2.13 (nex cassine) Loblolly bay 2.38 6.66 2.13 (Gordonia tasianthus) Swed bay 6.66 1.79 17.02 (Magnolia virginiana) Red maple 26.60 &#cer rubrum) Laurel oak 20.00 7.14 4.26 (Quercus laurvolia) Bull bay 7.14 (Magnolia grandijlord) Coastal plain willow 1.79 (Salix caroliniand) Wild olive 1.79 (9smanthus americand) I,oblolly pine 2.27 (pinus meda) Turkey oak 2.27 (Quercus laevis) ' M}{P=mesich=mm(*w-pah.Uen#,F'y BSW=' " ' I.-'-=Sk~st.ake; ASH= Ashley Laks, ROF=Ross Lake; G(*=Goose Lake; MCS=Mil (hek Swanp 220 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 38 PT. 11(8) Table 2. Relative density (%) of saplin[, in three habitat types on the Ordway Preserve Sites' MH MHP BSW Habitat Species SUG ASH ROS GOS MCS Sweet gum 30.95 12.50 1636 13.33 g.iquidambar styraciflua) Pignuthickory 19.05 1.82 7.32 (Sonya glabra) Water oak 16.67 43.75 30.91 46.34 11.11 (Quercus nigra) Laurel oak 14.29 18.75 12.73 2.44 6.67 Quercus laurifolia) Live oak 9.09 17.07 17.78 (!>terms virginiana) Black gum 7.14 9.75 2.22 (Nyssa sylvatica) Swamp bay 2.38 6.25 7.27 2.43 22.22 (Persea palustris) Longleafpine 2.38 4.44 (Pimapolustris) Sweet bay 2.38 1.82 7.31 15.56 (Magnolia virginiana) Dahoon holly 2.38 3 .64 4~1 6.67 (flex cassine) Carolina holly 1.82 (IIcx ambigua) Loblolly bay 2.38 1.82 2.44 (Gordonia lasianthu4 Red maple 12.50 9.75 0cer mbrum) Red bay 16.25 1.82 (Persea barbonia) Wild olive 3.64 4.88 (Osmanthus americana) Bullbay 3.64 (Magnolia grandiflora) Coastal plain willow 1.82 (Salix carliniand) Walking stick 1.82 (Aralia spinosa) 1 MHP-rnes~ hmmnock-saw-pahn* Lmdesstory, MH= ' ' ' I . ' y BSW=. -/-"==S14EsLake; ASH= Ashby Lake, ROS=Ross Laks, GOS=Goose Latz; MCS=Mit[ Cred~ Swamp. HAYSMITH: NEOTOA£4 F FLORIDANA IN NORTH-CENTRAL FLORIDA 221 Table 3. Relative density (%) ofseedling~ in three habitat types on the Ordwny Prese,ve Sites' MH MHP BSW Habitat Species SUG ASH ROS GOS MCS I.aurel oak 38.46 6.25 2.50 23.21 4.88 (Querms laurifolia) Water oak 30.77 56.25 62.50 32.14 9.76 (Quercus mgra) Live oak 17.50 14.29 21.95 (Quercus virginiand) Sweet gum 12.82 3.57 (Uquidambar styraciflud) Black gum 7.69 5.00 (Nyssa sylmtica) S,panlp bay 5.13 18.75 5.00 25.00 36.59 (Persea palsutris) Red bay 2.56 12.50 2.50 1.79 (Persea borbonia) Pignut hickory 2.56 (Carya glabra) L~ngleafpine 6.25 2.50 (Pinuspalustris) Red maple 2.50 (Acer mbrum) Loblolly bay 2.50 (Gordonia lasianthus) Sweet bay 21.95 (Magnolia virginiana) Dahoon holly 4.88 Ula cassine) 1 MHF '' ' ' y BS~¥=boUornhrMi kdv/mci swarrv, SUCANJ# Lake, ASH= A,hl~y L,kE, ROS=Ro= Lal# GOS=0003* L,ko; MCS=Mm Cm,* Sw~. 222 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 38 PT. II(8) Table 4. Relative density (%) ofshrubs in three habitat types on the Ordway Preserve Sites' MH MHP BSW Habitat Species SUG ASH GOS ROS MCS Saw-palmetto 64.28 44.44 31.25 (Serenoa repens) Wax myrtle 21.43 57.14 4.76 16.67 0*rica cenfera) Fetterbush 16.67 42.85 2.38 24.07 54.17 (Lyonia lucida) Beautyberry 23.81 (Callicarpa amencand) Butionbush 9.52 5.56 2.08 (Cephalant}nis occidentalis) Virginia willow 9.52 238 (Rea viginica) Cornmon pet~nmon 9.52 (D*spyros virginiana) Highbush blueberry 7.14 1.85 2.08 (Maccinium corybomm) Southern dewixny 2.38 (Rubus mvialis) Partridge berry 2.38 (Mitchella repeni) Gallberry 2.38 1.85 (nex glabra) Roseniaiy 4.76 (Ceratiola ericoides) Stagge~ush 2.38 6.25 (Lyonia fernigined) Tree sparkleberry 5.56 (Mawinium arboreum) Deerb~ry 4.17 (Macamum stamineum) 1 MHP=Inesicl~nock-6aw-painettour~[*Ny. MH= '' ' y BSW=2 .. '- -=S~ Latz; ASH= A«ey Lal< ROS=Ross Lak GOS=Goose I.Bir< MCS'*lin Cheir Swmip. HAYSMITH: NEOTOMA F. FLORIDANA IN NORTH-CENTRAL FLORIDA 223 Table 5. Relative density (%0 ofheits in three habitat types on the Ordway Preserve. Sites' MH MHP BSW Habitat Species SUG ASH ROS GOS MCS Fringed panicum 32.56 30.00 66.66 12.90 (Panicum ciliatum) Virginia chain fern 6.98 60.00 11.29 23.40 (Woochpardia virginica) Lizard tail 2.32 10.00 5.55 3.22 (Saunints cenmus) Dymorphic chainfern 6.98 10.64 (Woodwardia aerolata) Saw grass 11.11 (Stadium jamaicense) Maidencane 11.63 5.55 8.06 (Fanicum hemitomon) Cinnamon fan 5.55 1.61 10.64 (Osmunda cinnamomea) Sed& Fp 5.55 50.00 Elerhaa fod 13.95 07ephantopus datus) Green arum 6.98 6.38 (Pelt,mdra sve·) Virginia creeper 6.97 (Farthenocism,s quinquefolia) Brackenfern 8.06 29.79 (Fteridium aquitinum) Resurrection fem 1.61 (P*podiump*podioides) PritniOSe-Willow 1.61 g.,udwigia spp) Royal km 19.15 (Osmunda regatis) 1 MHP=methammock-saw-pelmettounderstory;MH= '' ' ' y BSW=boaornhandkdwoodswamp. SUC)=Suggst.ake, ASH= A*ley Lake; ROS=Ross Lake; GOSpes on the OrAny Presen,e. Sites' MH MHI' BSW Habitat Species SUG ASH ROS GOS MCS Bamboo vine 4.88 33.33 5.00 3.08 63.16 (Smil= laurifolia) Wild bamboo 41.46 20.00 2.50 1.54 10.53 (Smilar aunculata) Wild sarsaparilla 4.88 40.00 45.00 53.85 13.16 (Smitarglauca) Smilax pumilo 5.00 1.54 (Sarsapan#a vine) Hi#ush blackbeny 6.66 2.50 (Rubus argutus) Wild muscadine grape 43.90 35.00 32.31 13.16 (Mtisrotundfoha) Crossvine 2.44 2.50 (Bignonia capreolata) Groundnut 2.50 7.70 Opios americand) Yellowjessamin 2.44 (Gelsemium sempervirens) Ml{P= '. , ' " ' ' I .- '=/Ils. hammolk<%.n unddat"y BSW=' " * ~ , SUC)=Sig~s Lakr, ASH= Ashly I.atz; ROS=Roes Latz; GOS=Goose Lakz; MCS=Min Oeck Sminp. Stem density, also referred to as point-to-plant distance, was the most important measurement for estimating horizontal complexity. An ANOVA indicated significant differences in stem density by habitat for seedlings, shrubs, saplings, herbs, and vines, excluding only trees (p=0.05). Stem density was highest for saplings and seedlings in the mesic hardwood hammock with open understory (Suggs and Ashley lakes), and highest for shrubs in the bottomland swamp. A Kruskal-Wallis test demonstrated significant differences between habitat types (Zar 1984) for sapling, seedling, and shrub species but not for trees or herbs; herbs were the only factor excluded from the ANOVA results. After determining that significant differences were found among these strata, I conducted a multiple comparisons non-parametric test to indicate in which habitat point-to-plant distances were significant. This demonstrated that sapling and HAYSMITH: NEOTOMA F. PLORIDANA IN NORTH-CENTRAL FLORIDA 225 Table 7. Mean vegetation' heights (in mm) categorized by habitat type. Plant N Mean SE Trees MHP 109 10.65 0.48 MH 57 11.87 0.67 BSW 47 11.16 0.73 Saplingt MHP 108 3.10 0.29 MH 59 2.75 0.26 BSW 44 4.50 0.30 See~ings MHP 107 0.28 0.04 MH 60 0.36 0.05 BSW 41 0.31 O.06 Shnibs MHP 105 1.76 0.11 MH 55 1.13 0.16 BSW 47 1.59 0.17 Herbs MHP 79 1.01 0.48 MH 50 0.31 0.61 BSW 47 0.43 0.63 1 M}{12=inesic *mmock-saw·pahrwk unrilistory, MH=rnes~ 6,mz,ock-open wvdaitory, BSW=bottornksnd haid,wood swamp seedling stem density was highest in mesic hardwood habitats, but shrub stem density was highest in the bottomland swamp. Data on crown diameter also were analyzed using an ANOVA, and the Kruskal-Wallis and non-parametric multiple comparisons tests (p=0.05). The ANOVA indicated significant differences for saplings and shrubs by habitats, with the largest crown diameter for saplings, shrubs, and herbs, in the bottomland swamp. Comparison of the parametric and non-parametric analyses demonstrate that sapling and shrub crown diameter is largest in the bottomland swamp. This comparison also demonstrates that herb crown diameter is largest in the mesic hammock-saw-palmetto, and the bottomland swamp. Vegetation cover was assumed to be high in the mesic hammock with saw- palmetto understory, due to the dominance of saw-palmetto. However, stem density of saw-palmetto is low, and when viewed from an aerial position, shrub cover is consequently patchy. These factors were difficult to quantify using chosen vegetation sampling methods. It appears that by combining the evaluation for stem density with vegetation cover a more representative depiction of vegetation cover and consequent habitat quality for small mammals, such as woodrats, can be obtained. 226 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 38 PT. Il(8) Table 8. Mean values (in cm) of diameter basal height (DBH) ofvegelation' categorized byhabitaL Variable N Mean SE Tree MHP 110 23.9 1.6 MH 57 25.8 2.2 BSW 46 17.2 2.4 MHP 107 2.9 0.19 MH 59 2.6 0.25 BSW 45 4.0 0.29 Shrub MHP 105 5.6 0.62 MH 56 2.4 0.85 BSW 42 4.9 0.98 Herb MHP 80 1.2 0.22 MH 52 0.6 0.28 BSW 46 0.5 0.30 VIne MHP 105 0.7 0. 11 MH 58 0.6 0.15 BSW 36 0.9 0.19 1 MHP=mez hammodr.aw-patnetto und/.sto:y, MH=mes~ 1%171.(dien wd=Mory, BSW=bottornhnd tini„ood swamp. Population Variables Trap success varied per study site, but the mean for all study sites was 0.07, with 2755 total trap nights. Of the 218 captures, 78% (n=164) were recaptures (Table 9). Total animals captured at all study sites consisted of 32 males and 22 females, which resulted in a sex ratio of 1.6:1, favoring males. Weights of adult woodrats fluctuated considerably throughout the study period; a phenomenon also observed by Fitch and Rainey (1956). A Mest demonstrated that although male woodrats weighed significantly more than females, their body measurements were not larger (p=0.05). (The mean values for weight and body measurements are presented in Table 10.) Reproduction occurred throughout the year. A Chi-square test demonstrated that no significant differences existed in reproductive activity for males or females between months or seasons Or=0.05). However, peaks occurred in April and November for males, and July and November for females (Fig. 2). HAYSMITH: NEOTOMA F. PLORIDANA IN NORTH-CENTRAL FLORIDA 227 Table 9. Capture number and success rate per study site. mudy sites' ROS GOS SUG ASH MCS Total 75 46 39 6 52 3112 11 10 14 2 17 Trap success rate .08 .05 .07 .02 .11 1 ROS-Ross Late; GOS'nia atcida) Oak log 9 24.3 (Quercus spp.) Saw-palmetto 8 21.6 (Serenoa repens) Swamp bay 2 5.4 (Persea palustris) Water oak 2 5.4 Quercus n,gra) Live oak 1 2.7 (Quercus virginiana) Black pm 1 2.7 (Nyssa xylvatica) Wax myrtle 1 2.7 04*ca cerfera) Sweet gum 1 2.7 (Ijquidambar styrac*a) 18blolly pine 1 27 (Pinus taeda) Table 17. I ' 0 =0.051 Habitati Proportion Proportion Bon-ferroni Selection Available Utilized C.L Ross and Goose lakes: C oo rn n > 0.36 0.54 0.44 50an in height E = Ringe (black gum and fatterbush/wax m>Ttle); S = bayhead swamp; B = hardwood MH vegetation < 50 cm in height; P = proportion of utilization; G = grasses. 236 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 38 PT. Il(8) Microhabitat Use The fluorescent pigment body-dusting and capsule techniques were not highly successful. Fluorescent trails from body dusted animals were difficult to follow beyond a 4-m radius of the den. Trails that could be located were found in leaf litter, fetterbush, and on logs. Two animals whose dens were located in structurally complex vegetation provided the most important data. Pigment trails were located on a multitude of branches and stems at various heights. At one male den, 4720 cm of fluorescent trails were measured; 84.6% were arboreal and 15.4% were terrestrial. At one female den, I measured 5837 cm of trails; 75.1% were arboreal and 24.9% were terrestrial. DISCUSSION Habitat Woodrats occupy an array of habitats across their range. It is clear that northern populations of the eastern woodrat prefer rocky terrain. However, in the southeastern coastal plain, where these habitats are not abundant, Neotoma foridana are found in various woodlands. It is not clear if specific habitats are preferred among forest types. Barbour and Humphrey (1982) found the highest density in mature hardwood forests in Key Largo, Florida. Chamberlain (1928) recorded Neotoma floridana in wooded swampy areas in South Carolina. Pearson (1952) documented high densities of woodrats along mesic forest ecotones, but trapped them in swamps and high and low mesic hammocks in Gulf hammocks of Florida. This study confirms these researchers' reports, documenting woodrats in hardwood forests and swamps. However, in this study woodrats were strongly associated with three mesic forest habitats on the Ordway Preserve: hardwood hammock with saw-palmetto understory, hardwood hammock with open understory, and bottomland hardwood swamp. I documented similar plant species composition and relative densities of vegetation in the three mesic habitats occupied by woodrats. A notable feature of species composition is the species richness in all strata and in all habitats. However, significant differences were observed in vegetation structure on horizontal and vertical planes. The shrub stratum, followed by sapling and seedling strata, are probably the most significant to woodrat foraging and movement activities. Stem density was highest for shrubs in the bottomland swamp and for saplings and seedlings in the mesic hardwood habitat. The dense shrub stratum in the bottomland hardwood swamp results in an impenetrable wall of ericaceous shrubs. This attribute of structural complexity in the bottomland HAYSMITH: NEOTOMA F. PLORIDANA IN NORTH-CENTRAL FLORIDA 237 swamp habitat is probably the most significant vegetative component in all of the study sites. The bottomland swamp is also where densities remained the highest. This structural complexity can offer a woodrat protection from predators and higher food resource availability, particularly on a 3-dimensional scale. Population Variables Natural history characteristics documented in this study agree with observations by other researchers that woodrats are nocturnal (Fitch and Rainey 1956). Nocturnal behavior may be influenced by predation coupled with the longer daylight periods in the southern U. S., which would increase susceptibility to diurnal predators. Diurnal activity was infrequently recorded in previous studies, but diurnal observations by Schwartz and Schwartz (1959) and Rainey (1956) may be associated with shorter daylight periods, a dissimilar suite of predators, or different weather conditions. The sex ratio observed in this study was skewed towards males (1.6:1). Hersh (1981) documented a sex ratio of 1.2:1 on Key Largo woodrats. The skewed sex ratio towards males, may represent a population bias for males, or may be an anifact of female trap shyness. Reproduction of Neotoma f floridana occurred throughout the year in this study, which also was observed by Pearson (1952) in a study in Gulf Hammock, Florida. Pearson discovered reproductive peaks in the spring and fall, and peaks in this study occurred in April and November. Year-round reproduction in the southeastern U.S. is most likely influenced by climatic conditions, with warmer temperatures facilitating reproductive activities possibly due to higher availability of food resources. The social behavior of Neotoma is controversial , both social tolerance and intolerance have been reported in the literature. In this study woodrats were determined to be socially intolerant. They used solitary den sites, and rarely were observed in close proximity to each other. Social intolerance may be influenced by several factors including competition for food resources, shelter, and intrinsic population regulation. Kinsey (1976) reported seasonal differences in agonistic interactions between males and females of Neotoma that may have been associated with reproduction and rearing of young. However, no differences were observed in this study in seasonality and signs of social intolerance, which may have been actual or an artifact of sampling. Numerous intrinsic and extrinsic factors may influence the observed population fluctuations or declines of Neotoma f floridana on the Ordway Preserve. These may include social factors, resource scarcity, decreased reproduction, habitat differences, predation, or climatic factors. The population declines may be more profound due to the lower reproductive potential ofNeotoma, 238 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 38 PT. 11(8) compared to other small mammals. For example, high densities caused population declines in Rattus species (Davis 1953; Calhoun 1962), due to increased aggressive interactions, and decreased reproduction. The bottomland hardwood swamp consistently supported the highest densities of woodrats and may represent optimal habitat. The bottomland swamp is a large continuous tract of forest, and vegetation was the most structurally complex on horizontal and vertical planes. Plant species composition in the three habitats was very similar, but the mesic hammock with saw-palmetto understory and hardwood hammock with open understory were structurally less diverse. These habitats also are fragmented. Because of this fragmentation, woodrat densities may be restricted. The bottomland swamp may also serve as a 'source' to other connected but marginal habitats. Declines in density may also have been due to predation. The most significant amount of predation was attributed to barred owls (Strix varia) and great horned owls (Bubo virginiana). Snake predation was the second most commonly observed incidence of predation, attributed to the eastern diamondback rattlesnake (Crotalus adamanteus) and the yellow rat snake (Elaphe obsoleta). Of the incidences of predation, all known predation occurred to animals in the mesic hardwood hammock with open understory, and the mesic hardwood hammock with saw-palmetto understory. No predation was observed in the bottomland hardwood Mill Creek Swamp, which is dominated by extremely dense fetterbush that is impenetrable to large, avian predators, but presents no difficulty to reptilian predators. Timmerman (1989) and Franz (pers. comm.) suggested that reptile populations may be relatively abundant in the bottomland hardwood swamp. Wright (1989) considered owls to b. one of the top predators to opossum (Didelphis virginiana) in the mesic hardwood and xeric oak forests. If true for Neotoma, then near absence of owl predation in Mill Creek Swamp as observed in this study, could offer a partial explanation for the consistent higher densities found there. Conversely, the joint impact of snakes and owls in mesic forests, where both predators occur, may be partially responsible for the observed sharp fluctuations and declines in density. This study occurred in the middle of a severe drought in north-central Florida and may have been another factor influencing densities. During this period, Dodd (1992) documented a dramatic decline in densities of upland amphibians and reptiles on the Ordway Preserve, which was positively correlated to the reduction in rainfall. Additionally, Jones (1990) documented dramatic declines in Florida mice (Podomys floridanus) in the sandhill community which was positively correlated to declining rainfall levels. Woodrat densities at Goose and Ashley lakes crashed, and water levels in these two lakes dropped sharply. These two study sites also are isolated from other mesic forest types which could potentially limit woodrat dispersal. Negative effects from climatic changes such as drought on small mammal populations can result from numerous factors including a decrease in food supplies, HAYSMITH: NEOTOMA F. FLORIDANAIN NORTH-CENTRAL FLORIDA 239 lack of water, and other impacts on vegetation. Dependence by woodrats on drinking water for metabolic requirements is probably minimal; researchers have shown that Neotoma obtain their drinking water requirements primarily from vegetation (Linsdale and Tevis 1951; Dial 1988). Reduction in rainfall therefore, would likely impact woodrat populations primarily through vegetational changes. Neotoma f foridana are largely herbivorous in their feeding requirements. They consume a wide variety of plants (Voeten 1990), but dependence on specific plant products, such as oak mast, has been suggested by Murphy (1952) and Neal (1967) among other researchers. If long-term drought conditions reduce plant production such as fruiting and flowering, reduction in one or several major food resources could also influence woodrat densities. Resource scarcity can likewise affect reproduction which may be a critical factor for Neotoma because of its lower reproductive rate. Fitch and Rainey (1956) recorded declines in woodrat densities over a several year period. Neal (1967) documented sharp declines in a two year period. These researchers attributed the population declines to poor acorn crops and harsh winter conditions. Finally, secondary and compounded effects from declines in rainfall could affect plant resources and thus influence prey movement patterns as well as predator behavior (i.e. concentrations of predators in mesic habitats), which could influence densities of woodrat populations that live in these restricted mesic hammocks. Home Range and Activity Activity patterns of Neotoma observed in this study are similar to the activity periods observed by Wiley (1971), with high activity peaks between 2000 and 2200 h. Peaks of activity may be related to environmental variables, such as changes in temperature, precipitation, or light. Activity may also be influenced by foraging and gathering activities, related to the compulsive habit of woodrats to collect and store large quantities of food and other items in their dens (Worth 1950). Numerous factors may influence the home range size and movement patterns of woodrats such as: density fluctuations of woodrats in all study sites, year round reproductive activity, the presence or absence of horizontal and vertical vegetation complexity, and use of multiple dens. These factors most likely influence the high variability of home range sizes observed in this study. The importance of food resources to woodrat activities is apparent; while Neotomafloridana do not depend on single plant taxa, they are undoubtedly highly tuned to micro- and macro-distribution of food resources, plant phenologies, and environmental variables. Food resources likely influence woodrat movement patterns and home range size, but are extremely difficult to quantify. Fitch and Rainey (1956) documented changes in home ranges over time, attributing it to changes in food resources. Therefore, the variability in home range size over time 240 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 38 PT. II(8) documented in this study undoubtedly is linked to the diverse plant species composition in the forests and consequent production of diverse plant products. Male and female home range sizes were similar in this study, although females were slightly larger than males. This contrasts to other studies where male home range sizes were larger than those of females (Fitch and Rainey 1956; Neal 1967). These results may be influenced by sample size and the high within-sample variability of the radiotelemetry observations. Contrary to many other mammalian radiotelemetry studies, male and female home ranges did not overlap in most of the study sites. The three ranges that did overlap (n=23) were male-female, and may have been associated with reproductive periods. In the analyses of woodrat home ranges according to habitat type, size was on average smaller in the mesic hardwood and bottomland swamp habitats. In these areas densities were lowest although all sites had wide-ranging individuals. However, this relationship between home range and habitat type was statistically significant in this study. The influence of density on home range size documented by many researchers can result in a negative correlation between density and home range. Hence, when densities are low, home ranges frequently increase. That relationship was documented in this study, where individuals increased home ranges after the disappearance of other woodrats who had adjacent home ranges. Another factor possibly influencing home range size is the complex horizontal and vertical vegetation structure in the bottomland hardwood swamp and mesic hammock with open understory habitats (e.g. high stem density coupled with large crown diameter). This complex vegetation structure may influence animal home range size due to higher 3-dimensional space use in this study. In other words, arboreal habitat use increased when woodrat dens were located in structurally complex vegetation. In contrast, where vegetation structure is less complex on vertical and horizontal planes, home ranges may be larger due to less three dimensional space use: animals may be travelling out rather than up in their movements (Mel Sunquist, pers.comm.). This could contribute to a smaller linear home range size in less complex habitats. The bottomland swamp is also a large contiguous area. The other habitats, mesic hammock with saw-palmetto understory (MHP) at Ross and Goose lakes and mesic hardwood forest (MH) at Suggs and Ashley lakes, are narrow strips of mesic forest ranging from 50 m to 200 meters in width. Based on captured immigrants and disappearance of radio-collared animals, I suspect woodrats may make large scale movements. The Mill Creek Swamp populations remained the highest throughout the study, and the Ross and Suggs lakes populations declined but rebounded. These two areas are distally connected to Mill Creek Swamp, facilitating potential animal movements. The Ashley and Goose lakes sites are farther away and are disjunct from Mill Creek Swamp, potentially causing difficulties for woodrat dispersal into these areas, and populations disappeared in these two areas. HAYSMITH: NEOTOA£4 F. FLORIDANA IN NORTH-CENTRAL FLORIDA 241 Dens Large stick piles that are typical of Neotomafloridana den structures in other geographical areas were rare in this study. Most dens were subtemnean, which agrees with findings by Pearson (1952) in Florida. Among other factors, these could be related to plant substrate types available in the habitat or soil types. Dens were also associated with several plant species. High variability in plant species at dens may be indicative of the importance of numerous plant resources, or plant structure for making dens, rather than specific species composition in den site selection. Cover and complex vegetation structure at several of the dens were also high. Radiotelemetry in this study documented highly variable home ranges, but 54% were less than 0.5 ha. Rainey (1956) suggested woodrat home ranges were frequently in close proximity to the den. This vegetation complexity at dens would be advantageous to small ranging animals for protection and food resource availability. The most significant characteristic of den site location was that 69% of all dens were located on an ecotone. Access to diverse resources may therefore be an important influencing factor in den site selection. Several woodrats changed den sites, which may be attributed to preferred structural features at specific dens (i.e. stem density and cover), geographical location, or avoidance of parasites. Woodrat dens may be a limiting factor in the community. Many times woodrats relocated into a den after the previous occupant disappeared, and some dens were subsequently used by several different tenants, indicating there may be "preferred den sites." Dial (1988) reported "preferred den sites" also, indicating that dens may be in limited supply. Other studies have documented that mammals will evacuate dens when parasite loads become too high. Bot fly larvae infestation in woodrats was relatively high, and bot flies are known to lay their eggs at rodent dens. Therefore, this could account for some of the den infidelity observed. However, "preferred dens" may simply be more "functional" as houses than others. The use of one or more dens also influences home range size, particularly ifdens are not in close proximity. Habitat Utilization Utilization of micro-habitats by woodrats was proportional to the habitats available. As expected, they utilized all of the micro-habitat types available except bare ground. A preference was exhibited towards saw-palmetto in the mesic hammock-saw-palmetto habitat, and swamp for the animals inhabiting the edge of Mill Creek Swamp, which may be related to vegetation cover requirements. 242 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 38 PT. II(8) Woodrats were distributed in clusters in continuous habitats, forming what I entitled "woodrat neighborhoods," although they used solitary den sites. The reported "colonial" nature of woodrats in previous studies in Florida (Pearson 1952; Chamberlain 1928) may be factual, but also may be an interpretation of large scale "coloniality," or a cluster distribution, observed in this study. It is not clear whether these cluster distributions are associated with micro-habitats, although they do not appear to be associated with macro-habitat types because woodrats were not trapped throughout continuous habitats. While I had minimal success using fluorescent pigment powders to document three dimensional space use, the data in this study suggest the arboreal nature of Neotoma f floridana. The fluorescent pigment technique has been highly successful in some studies (Mullican 1988) to document home range and arboreal activity (Goodyear 1989). However, if vegetation density is high, and if the animal grooms compulsively, release of pigment powder can be significantly reduced. LITERATURE CITED Barbour, D.B., and S.R. Humphrey. 1982. Status and habitat ofthe Key Largo woodrat and cotton mouse (Neotoma floridana smalli and Peromyscus gossypinus allapaticola). 3. Mamin. 63:144-148. Burt, W.H. 1943. Territoriality and home range concepts as applied to mammals. J. Manm 24:346-352. Byers, C.R, and B.K. Steinhorst- 1984. Clarification ofa technique for analysis of utilization-availability data. J. Wildl. Mgmt 48: 1050-1053. Calhoun, J.B. 1962. The ecology and sociology ofthe Norway raL Washington, D.C., U.S. DepL Health Educ. Welfare, Public Health Serv. 288 pp. Chamberlain, E.B. 1928. The Floridawood rat in South Carolina J. Manim 9:152-153. Coleman„ J.S., and AB. Jones, III. 1988. User's guide to TELEMBB: Computer analysis system for radiotelemetry data. Dept Fish. Wildl., Virginia Polytechnic Inst and State Univ., Blacksburg VA. 49 pp Davis, D.E. 1933. The characteristics ofrat populations. Quart Rev. Biol 28:373-401. Dial, K.P. 1988. Three sympatric species of Neotoma: dietary specialization and coexistence Oecologia 76:531-537. Dodd, C.K. 1992. Biological diversity of a temporary pond herpetofauna in North Florida sandhills. Biodiv. Conserv. 1:125-142. Eisenberg, J.F. 1989. Mammalian species of the Ordway Preserve. Univ. Florida Florida Mus. Nat. Hist., Ordway Preserve Rept 1, Gainesville. 73 pp Fitch, H.S., and D.G. Rainey. 1956. Ecological observations on the woodrat Neotoma floridana. Univ. Kansas Publ.Mus. Nat. }list 8:499-333. Franz, R., and D.W. Hall. 1990. Vegetative communities and annotated plant lists for the Katharine Ordway Preserve-Swisher Memorial Sanctuary, Putnam County,Florida. Univ. Florida Florida Mus. Nat. Hist, Ordway Preserve RepL3, Gainesville. 89 pp, Goertz, J.W. 1970. An ecological study ofNeotoma floridana in Oklahoma. J. Mamm. 51:94-104 Goodyear, N.C. 1989. Studying fine scale habitat use in small mammals. J. Wildl. Mgmt 53:941-946. Greer, KE. 1978. Eastern woodrats in south-central Florida Florida Sci. 41:191-192. Hamilton, W.J.,Jr., and J.O. Whitaker. 1981. Mammals ofthe Eastern United States. Cornell Univ. Press, Ithaca, New York. pp 207-211. Hamilton, W.J.,Jr. 1953. Reproduction and young of the Florida woodrat Neotoma Ljloridana. J. Mamm. 34: 180-189. HAYSMITH: NEOTOMA F. PLORIDANA IN NORTH-CENTRAL FLORIDA 243 Harris, S., W.J. Cresswell, P.G. Forde, W.J. Trewhella. T.Wollard, and S. Wray. 1990. Home range analysis using radiotracking data: a review of problems and techniques particularly as applied to the study ofmammals. Mamm Rev. 20:97-123. Harvey, M.J., and R.W. Barbour. 1965. Home range of Microtus ochrogaster as determined by a modified minimum area method. J. Mamm. 46:398-402. Hays, R.L., C. Summers, and W. Seitz. 1981. Estimating wildlife habitat variables. Washington, D.C. U.S. Dept.Int, Fish Wildl. Serv. FSW/OBS-81/47. 111 pp. Hersh, S. 1981. Ecology ofthe Key Largo woodrat (Neotoma.#oridana smatti). J. Marna 62: 199-201. Humphrey, S.R. 1988. Density estimates of the endangered Key Largo woodrat and cotton mouse (Neotoma floridana smalli and Peromyscus gossypinus allapaticola) using Be nuited grid approach J. Mamm 69:524-531. Jones, C.A. 1990. Microhabilat use by Podomys floridanus in the high pine lands of Putnam County, Florida. Ph.D. Dissertation, Univ. Florida Gainesville. Kaufmann, J. H. 1962. Ecology and social behavior ofthe coati Nama narica on Barro Colorado Island, Panama. Univ. California Publ. Zoo!. 60:95-222. Kinsey, K.P. 1976. Social behavior in confined populations of the Allegheny woodrat Neotomafloridana magister. Anim. Behav. 24:181-187. Layne, J.N. 1974. The land mammals of South Florida. Pp.386-413 in P.J. Gleason, ed. Environments of South Florida: Present and past Mem Miami Geol. Soc. 2:1-452. L,eman, C.A, and P.W. Freeman. 1985. Tracking mammals with fluorescent pigments: A new technique. J. Mamm.66:134-136. Linsdale, J.M., and L.P. Tevis, Jr. 1951. The Dusky Footed Woodrat Univ. California Press, Berkeley. 664 pp. Ludwig, J.A, and J. F. Reynolds. 1988. Statistical Ecology. John Wiley and Sons, New York. 337 pp Michener, G.R. 1979. Spatial relationships and social organization of adult Richardson's ground squirrels. Canadian J. Zool. 57:125-139. Mullican, T.R. 1988. Radiotelemetry and fluorescent pigments: A comparison of techniques. J. Wildl. Mgmt 52:627-631. Murphy, M.F. 1952. Ecology and helminths of the Osage woodrat, Neotoma foridana osagensis, including descriptionof Longistriara neotoma n. sp. (Trichostrongylidae). Amer. Midi. Nat. 48:204-218. Neal, WA 1967. A study ofthe ecology of the woodrat m the hardwood forests of the lower Mississippi River Basin. M.S. Thesis, Inuisiana State Univ., Baton Rouge. 116 pp. Pearson, P.G. 1952. Observations concerning the life history and ecology of the wood rat Neotoma floridana/londana (Ord). J. Mamm. 33:459-463. Poole, E.L 1940. A life history sketch ofthe Allegheny woodraL J. Mamm. 21:249-270. Rainey, D.G. 1956. The Eastern woodrat life history and ecology. Univ. Kansas Mus. NaL Hist. 8:499-533. Schwartz, A., and E.P. Odum. 1957. The woodrats ofthe eastern United States. J. Manm 38:197-206. Schwartz, C.W., and E.R. Schwartz, 1939. The Wild Mammals of Missouri. Univ. Missouri Press, Columbia. 341 pp Sciascia, J. 1990. Eastern woodrats decline in tri-state area. Nongame News. New Jersey Dept. Envir. Prot, Hampton. 2 pp Sunquist, M.E. 1989. Comparison of spatial and temporal activity of red foxes and gray foxes in north-central Florida Florida Field Nat 17:11-18. Timmennan, W. 1989. Home range, habitat use, and behavior of the Eastern diamondback rattlesnake. M.S. Thesis. Univ. Florida Gainesville. Voeten, M.M. 1990. Aspects ofthe f6*ling ecology of the Eastern woodraL M.S. Thesis, Univ. Utrechi The Netherlands. 41 pp. Wiley, R.W. 1971. Activity periods and movements ofthe Eastern woodrat Southwestern Nat 16:43-54. Wiley, R. W. 1980. Mammalian species. Neotomafloridana. No. 139, pp. 1-7 Worth, C.B. 1950. Observations on the behavior and breeding ofcaptive rice rats and woodrats. J. Mamm. 31:421-426. Wright, D. 1989. Mortality and dispersal of juvenile opossums (Didelphis virginiana). M.S. Thesis, Univ. Florida Gainesville. Zar, J.H. 1984. Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, New Jersey. 718 pp