Habitat associations and diversity of the rodent community in the Edwards Plateau of Central Texas.
Goetze (1998) recently published an extensive treatment on the mammals of the Edwards Plateau of Central Texas. A field survey of terrestrial vertebrates conducted at Colorado Bend State Park (CBSP; San Saba and Lampasas counties) provided an opportunity to contribute further to the knowledge of small-mammal ecology within this region. The park lies within the Llano Uplift region at the northeastern extent of the Edwards Plateau (Goetze 1998) and contains a variety of geographic features, including the Colorado River, its associated creeks and canyons, caves, sinkholes and limestone outcrops, that provide a variety of habitats utilized by many species of mammals. The mammalian fauna found here, in the Balconian Biotic Province, includes no endemic forms, but is a mixture of Austroriparian, Tamaulipan, Chihuahuan and Kansan species (Blair 1950) and represents an east-west transition zone for terrestrial vertebrates (Gehlbach 1991).
This project entailed field study (1) to determine relative densities of small mammals (rodents) in an array of terrestrial habitats, (2) to assess vegetative cover in these habitats and (3) to relate diversity and abundance of small mammals to vegetative cover. Study findings were used to examine habitat preferences of small mammals and were evaluated in terms of biogeographic and management considerations.
MATERIALS AND METHODS
Study site.--Colorado Bend State Park is situated along the Colorado River 10 km south of Bend, Texas, on County Road 257 between the cities of Lampasas (29 km northeast) and San Saba (29 km northwest). The park comprises 2,157 ha separated by the river into two sections, 245 ha east of the river in Lampasas County and 1,912 ha west of the river in San Saba County. Ordovician limestones and dolomites underlie the park. Groundwater erosion has created a system of caves, sinkholes and collapse zones in the park. Mollisols from the Tarrant, Kavett and Tobosa soil series units are found in the area of CBSP (Goetze 1998). The primary soil types found within the park include the Roughcreek-Eckrant (shallow and very shallow, gently sloping to rolling, very stony and stony) and Yahola (deep loamy soils, frequently flooded) series (U.S.D.A. 1982).
CBSP is located in Blair's (1950) oak/hickory/mesquite vegetation region, and it lies toward the eastern edge of the Edwards Plateau vegetation zone (Gould 1975). Important climax grasses of the Plateau include switchgrass (Panicum virgatum), gramas (Bouteloua sp.), Indiangrass (Sorghastrum nutans), wildrye (Elymus sp.), curly mesquite (Hilaria sp.), buffalograss (Buchloe dactyloides) and several species of bluestems (Schizachyrium, Bothriochloa, Dichanthium and Andropogon) (Gould 1975). Rough rocky areas are common throughout the Plateau and usually support tall or mid-grass understory and a brush overstory of Plateau live oak (Quercus fusiformis), shinnery oak (Quercus havardii), Ashe juniper (Juniperus ashei) and mesquite (Prosopis glandulosa).
Field methods.--Permanent sampling transects were established in each of eight major habitat types in CBSP as described in a habitat map provided by Texas Parks and Wildlife Department: two in terrace grasslands, two in riparian woodlands, three in juniper/oak/persimmon mosaic habitats, four in live oak parks, two in oak woodlands/juniper shrub habitats, one in juniper shrub habitat, one in juniper woodlands, and three in rangeland/rock outcrop habitat. Irregularly sampled transects also were established in all habitats in the park, including four in mesquite parks habitat.
Rodents were trapped along transects of 30 Sherman live-traps set in approximately a straight line approximately five meters apart. Traps were baited with crimped oats and sunflower seeds, and left overnight. Permanent transects always were set for two consecutive nights. Rodent sampling was conducted a total of 31 nights from May 2001 to April 2002. Most permanent transects were sampled for a total of 240 trapnights. Relative densities of rodent species were measured as individuals per 100 trapnights. Most rodents were released at the site of capture, but at least one voucher specimen of each species was kept for each transect and deposited into the Collection of Vertebrates in the Biology Department, Baylor University in Waco, Texas.
Vegetation surveys were conducted along the 18 permanent transects at three month intervals from May 2001 to April 2002 to obtain quantitative information on vegetative cover. Vegetation sampling generally followed methods established by Daubenmire (1959), a variation of which has been used successfully in previous studies of small-mammal habitat associations (Hanchey & Wilkins 1998; Wilkins 1995). A vegetation sampling station was established one meter from every 5th, 10th, 15th, 20th and 25th station of rodent trapping transects. These stations consisted of a 0.5 m by 0.2 m quadrat, a total area of 0.10 [m.sup.2]. Cover was evaluated for the following vegetation classes: dead herbaceous material, forbs, Johnson grass (Sorghum halepense), rye grass (Lolium perenne), eastern gammagrass (Tripsacum dactyloides), switchgrass (Panicum virgatum), other grasses, cactus (predominantly prickly pear, Opuntia lindheimeri) and woody vegetation. Cover was assessed at multiple heights above the soil surface (Wilkins 1995): dead material was measured at 5, 10, 25 and 50 cm above the ground, forbs at 5, 10 and 25 cm, and all grasses and cactus at 10, 25 and 50 cm. Woody vegetation was estimated at one meter and above.
Each vegetation class was assigned a value from 0 to 6 based on the estimated amount of ground cover, 0 = 0% cover, 1 = 1-5% cover, 2 = 6-25% cover, 3 = 26-50% cover, 4 = 51-75% cover, 5 = 76-95% cover and 6 = 96-100% cover. The values used in the computations are the midpoints of each of the percentage ranges: code 0, 0%; 1, 3 %; 2, 15.5%; 3, 38%; 4, 63%; 5, 85.5%; 6, 98%.
Analytical methods.--Descriptive statistics for vegetative cover were computed using the SAS statistical program Procedure MEANS (SAS Institute 1985). The percent cover for each vegetation category for each of the five sampling stations was used to calculate the mean vegetation cover for each transect. The means of all transects in a habitat then were used to calculate the mean cover for each vegetation category for that habitat. The habitat means are, therefore, the means of transect means.
For each rodent species and for all rodents combined, the mean relative density for each transect was calculated by dividing the number of rodents caught on a given transect during the study by the number of trapnights, then multiplying by 100, to yield the number of individuals caught per 100 trapnights. The mean relative density for each habitat then was calculated by averaging the means of all transects in that habitat. To evaluate possible habitat affinities, rodent densities were correlated with amount of cover provided by each of 23 categories of vegetation by using Procedure CORR of the SAS statistical program (SAS Institute 1985) to calculate Pearson's correlation coefficient, r; the correlation was judged to be significant at P [less than or equal to] 0.05. The Shannon-Wiener diversity index for finite populations, H', was calculated for each habitat using the DIVERS program in Bio[SIGMA]tat, SigmaSoft software (Pimentel 1993).
RESULTS AND DISCUSSION
Vegetative description of habitats.--Substantial vegetative cover in the juniper/oak/persimmon habitat was provided by the other grasses categories (39% summed for all three levels) consisting of Texas winter grass (Nasella leuchotrichia), mesquite grass (Bouteloua rigidiseta), spreading love grass (Eragrostis pectinacea), King Ranch bluestem (Bothriochloa ischaemum), arrowfeather (Aristida purpurea), love grass (Eragrostis secundiflora), fox grass (Setaria leucophila) and common switchgrass (Panicum capillare) (Table 1, Fig. 1). Woody vegetation (Ashe juniper shrubs and trees, Plateau live oaks and Texas persimmon [Diospyros texensis]) averaged 23.9% cover, making it the next most abundant type of cover. The remaining cover was provided by dead herbaceous material, forbs and cactus.
[FIGURE 1 OMITTED]
The dominant cover type in the juniper shrub habitat was woody (juniper shrubs, with some persimmon and live oak), averaging 33.7% (Table 1, Fig. 1). The most abundant herbaceous cover was the other grasses category (totaling approximately 23%), including wooly grass (Eroneuron pilosa), starved rosette grass (Panicum depaperatum), bent grass (Eragrostis hyemalis), oldfield threeawn (Aristida oligantha) and threeawn (A. wrightii). Other vegetation provided only sparse cover. Forbs, dead material and cactus were minor contributors, totaling approximately 26%.
The predominant source of cover in the juniper woods was juniper trees; woody vegetation averaged 51% coverage (Table 1, Fig. 1). Other types of vegetation were sparse, all grasses, forbs and dead material totaling less than 25 % coverage. Rock crevices and outcrops of limestone were common in this habitat.
The live oak parks habitat was characterized by considerable amounts of dead material at 5 cm (approximately 30%), but woody vegetation (live oak, some juniper and persimmon) provided the most cover of all vegetation categories (34%; Table 1, Fig. 1). Other grasses were the most abundant herbaceous material (35%) including winter grass, spear grass, oldfield threeawn, love grass, ovalleaf sedge (Carex cephalophora), arrowfeather, starved rosette grass, and hairy gramagrass (Bouteloua hirsuta).
Woody vegetation consisting of Plateau live oak, Ashe juniper and Texas persimmon provided the greatest amount of cover in the oak woodland/juniper shrub habitat (33%; Table 1, Fig. 1). Dead material was a major component of the habitat at 5 cm (23%), and other grasses were the most abundant herbaceous material (approximately 22%). These grasses include windmill grass (Chloris verticilata), arrowfeather, brome grass (Bromus japonicus), flat sedge (Cyperus stirgosus), starved rosette grass, love grass, wooly grass (Eroneuron pilosa) and common curlymesquite (Hilaria belangeri).
Forbs (approximately 43%) and other grasses (approximately 39%) were the major vegetative components of the rangeland/rock outcrop habitat (Table 1, Fig. 1). Other grasses consisted of winter grass, spear grass, oldfield threeawn and wild canary grass (Phalaris caroliniana). Woody vegetation (4%; juniper, oak, persimmon) and other categories were relatively minor components of this habitat.
The riparian woodlands follow the courses of the Colorado River and creeks within the park. Woody plants (Plateau live oak, red oak [Quercus rubra], black walnut [Juglans nigra], pecan [Carya illinoisensis], red ash [Fraxinus pennsylvanica], American elm [Ulmus americana]) was the main vegetative component (46%; Table 1, Fig. 1). However, rock outcrops and ledges occurred extensively, making exposed rock surfaces a significant component of this habitat. Dead material was abundant at 5 cm, averaging 34.7% coverage, with all other vegetation categories as only minor contributors to cover.
The terrace grassland is characterized by a diversity of grasses, together yielding the densest grassy cover in the park. Summed for all sampling heights, the other grasses category provided approximately 109% coverage (Table 1, Fig. 1). These grasses include brome grass (Bromus catharticus), crab grass (Digitaria sanguinalis), hairy seed paspalum (Paspalum pubiflorum), love grass, spreading love grass (E. pectinacea) and wild rye (Elymus virginicus). Several grasses present in this habitat, including switchgrass (approximately 1.5% cover), Johnson grass (approximately 3.5%) and rye grass (approximately 19%), were not found in other sampled habitats of the park. Woody vegetation, consisting of oak, ash, cedar elm, pecan, walnut and juniper, provided a limited amount of cover (approximately 13%) in this habitat.
Rodent community.--Six species of rodents (216 individuals) were trapped during the study (Table 2): Peromyscus pectoralis (white-ankled mouse, n = 180), Sigmodon hispidus (hispid cotton rat, n = 29), Chaetodipus hispidus (hispid pocket mouse, n = 3), Reithrodontomys fulvescens (fulvous harvest mouse, n = 2), Peromyscus maniculatus (deer mouse, n = 1), and Baiomys taylori (pygmy mouse, n = 1). Total sampling effort was 5,940 trapnights. Overall relative density of the rodent community was 3.64 individuals per 100 trapnights.
For three habitats, Peromyscus pectoralis was the only rodent species captured. Peromyscus pectoralis was least abundant in the juniper woods (14 individuals; 1.94 per 100 trapnights) and was most abundant in the riparian woodlands (39 individuals; 7.03 per 100 trapnights). This species reached intermediate densities (25 individuals; 2.67 per 100 trapnights) in the juniper shrub habitat. Because only one species of rodent was caught, the Shannon-Wiener diversity index was 0 for all three habitats.
In four other habitats (juniper/oak/persimmon, oak woodland/juniper shrub, live oak parks and mesquite parks), only Peromyscus pectoralis and Sigmodon hispidus were captured (Table 2). In all four situations, P. pectoralis was the more-abundant species. Relative densities for P. pectoralis ranged from 1.56 (live oaks parks) to 2.83 (oak woodland/ juniper shrub) individuals per 100 trapnights. For Sigmodon hispidus, densities ranged from 0.05 (juniper/oak/persimmon mosaic and live oaks parks) to 0.42 (rangeland/rock outcrop) individuals per 100 trapnights. Shannon-Wiener diversity indices for these habitats were intermediate to the other habitats sampled at the park: juniper/oak/persimmon, 0.21; live oak parks, 0.47; oak woodland/juniper shrub, 0.57; and mesquite parks, 0.94.
Rangeland/rock outcrop transects yielded three species of rodents, Peromyscus pectoralis, Sigmodon hispidus and Reithrodontomys fulvescens (Table 2). Peromyscus pectoralis was the most prevalent species with 17 individuals captured (mean of 1.77 per 100 trapnights). Sigmodon hispidus was uncommon in this habitat with only four individuals captured (mean of 0.42 per 100 trapnights). Reithrodontomys fulvescens was rare in this habitat as only one individual was captured (in July 2001) for a mean of 0.10 individuals per 100 trapnights. The Shannon-Wiener diversity index for the rangeland/rock outcrop habitat was 0.94.
The terrace grassland habitat supported the most diverse rodent community in the park (Table 1), producing six species. The most abundant species was Peromyscus pectoralis with 23 individuals captured (mean of 2.06 individuals per 100 trapnights) for all sampling periods. Sigmodon hispidus was more abundant in this habitat than in any of the other habitats sampled (15 individuals; mean of 1.10 per 100 trapnights). Reithrodontomys fulvescens was uncommon in this habitat, only one individual captured (mean of 0.04 per 100 trapnights). Three species were found only in the terrace grassland habitat: Three Chaetodipus hispidus were captured (mean of 0.14 per 100 trapnights) in August, October, and November 2001, in the soft soil along the top of the Colorado River canyon. One Peromyscus maniculatus was captured (mean of 0.05 per 100 trapnights). One Baiomys taylori was captured (mean of 0.15 per 100 trapnights). The Shannon-Wiener diversity index for the terrace grassland was 1.66.
Rodent habitat affinities.--Correlation analyses were conducted to determine affinities of the three most abundant rodent species (Peromyscus pectoralis, Sigmodon hispidus and Reithrodontomys fulvescens) encountered on the permanent transects for which vegetative cover was assessed. Peromyscus pectoralis, the most abundant rodent species in all habitats, was not significantly correlated with any of the vegetation categories (Table 3). Peromyscus pectoralis commonly occupies habitats containing rocky (limestone) slopes (Baccus & Horton 1984; Etheredge et al. 1989). However, P. pectoralis was present in all habitats at CBSP, and though it reached its highest density in the riparian woodlands, which had abundant slopes and rock outcrops, it also was common in the mesquite parks and other habitats which offer little to no exposed rock. Such generalist tendencies reflect the findings of Hanchey & Wilkins (1998) in which P. pectoralis was common in a variety of habitats in north-central Texas (Johnson County).
Unlike P. pectoralis, both S. hispidus and R. fulvescens showed significant correlations to many vegetation categories (Table 3). Both species correlated positively with switchgrass, Johnson grass, rye grass and the other-grasses category at all measured heights. Switchgrass, Johnson grass and rye grass occurred only in the terrace grassland habitat, which was one of only two habitats in which R. fulvescens was captured, and was where S. hispidus was most plentiful. Reithrodontomys fulvescens showed a significant negative correlation with woody vegetative cover at and above one meter, a category whose lowest values were in terrace grassland and rangeland/rock outcrop habitats, the other habitat in which R. fulvescens was caught. These findings substantiate previous findings in which S. hispidus was found to prefer grassy habitats over wooded areas (Hanchey & Wilkins 1998; Goetze 1998). These findings also agree with previously documented preferences of R. fulvescens for grassy areas and avoidance of woody areas (Hanchey & Wilkins 1998; Davis & Schmidly 1994).
Biogeographic considerations.--In his recent treatise, Goetze (1998) documented nine species of rodents in San Saba County. This current study verified four of these at CBSP: Reithrodontomys fulvescens, Peromyscus pectoralis and Baiomys taylori via trapping and Sciurus niger (fox squirrel) via sighting and a dead-on-road specimen. Bailey's map (1905: Fig. 10) included San Saba County in the distribution of the fox squirrel. However, the distribution he depicted for the pygmy mouse was restricted to south Texas and the Gulf Coastal Plain (Bailey 1905: Fig. 15). A century ago, the pygmy mouse had not yet invaded the Edwards Plateau, whereas now (Sehmidly 2002: Map 14) this species is absent only from the Trans-Pecos and from northeastern Texas, broadly defined. Range expansion of this neotropical species has been remarkably rapid.
Sampling during this study did not reveal four additional species (Spermophilus variegatus, rock squirrel; Peromycus attwateri, Texas mouse; Peromyscus leucopus, white-footed mouse; Neotoma micropus, southern Plains woodrat) that Goetze (1998) listed for San Saba County, although seemingly suitable habitats were present. No evidence of pocket gophers was seen at CBSP where suitable sandy soils are present discontinuously only within the narrow corridor of the Colorado River. However, only a short distance (10-15 km) upstream of CBSP, near the town of Bend, mounds of pocket gophers (Geomys texensis) regularly were noted in sandy soil along highway roadsides. Bailey (1905) noted the distribution of pocket gophers in the Edwards Plateau to be associated with sandy soils of river valleys, including the Colorado River. The other three species of rodents that were captured at CBSP (Sigmodon hispidus, Peromyscus maniculatus, Chaetodipus hispidus) are known widely throughout Texas (Davis & Schmidly 1994), but Goetze (1998) found no prior records of them for San Saba County. Bailey's (1905: Fig. 17) illustration of the range for cotton rats positioned San Saba County in the transition between the eastern and western subspecies of Sigmodon hispidus.
Management implications.--This research has evaluated the small-mammal community in the array of terrestrial habitats recognized at Colorado Bend State Park. Species richness of the rodent community differs across these habitats. Greater mammal richness accompanies greater habitat diversity as manifest in the number and type of categories of vegetation providing cover. Habitats dominated by woody vegetation tended to support fewer species of small mammals. Presence of herbaceous ground cover, with reduction of woody cover, corresponded to a greater number of species of small mammals.
For a land manager striving to maintain or restore high levels of biodiversity of small mammals, it is evident that the most important terrestrial habitat at this site is the terrace grasslands. This habitat supports all six species of rodents captured at the park. Rodent species caught in all other habitats represent subsets of those in the terrace grasslands. In area, the terrace grasslands are the least extensive of all terrestrial habitats in the park (except for the mesquite-parks habitat), occurring sporadically in narrow bands along the river frontage. This habitat, however, also is a primary one in which human recreational activities are concentrated. These competing interests might pose a management challenge, though approached appropriately, this might present an interpretative opportunity leading to a greater public appreciation of mammalian diversity.
Table 1. Mean percent cover (followed by standard deviation) provided by each of 23 vegetation categories for each habitat at Colorado Bend State Park, San Saba and Lampasas counties, Texas. Habitat means based on the means of the permanent transects within the habitat. The habitats include juniper/oak/persimmon mosaic (JOP), juniper shrub (JS), juniper woods (JW), live oak parks (LOP), oak woodland/juniper shrub (OWJS), rangeland/rock outcrop (RRO), riparian woodland (RW), and terrace grassland (TG). Habitat Vegetation category JOP JS JW Dead, 5 cm 7.72, 3.88 5.7, -- 4.43, -- Dead, 10 cm 1.9, 0.46 2.45, -- 0.45, -- Dead, 25 cm 0.2, 0.35 0, -- 0, -- Dead, 50 cm 0.07, 0.12 0, -- 0, -- Forbs, 5 cm 10.61, 6.57 7.85, -- 2.58, -- Forbs, 10 cm 8.94, 7.03 7.25, -- 2.75, -- Forbs, 25 cm 4.19, 3.96 0.3, -- 0.05, -- Cactus, 10 cm 1.18, 1.04 1.7, -- 0, -- Cactus, 25 cm 0.53, 0.61 1.55, -- 0, -- Cactus, 50 cm 0, 0 0.15, -- 0, -- Switchgrass, 10 cm 0, 0 0, -- 0, -- Switchgrass, 25 cm 0, 0 0, -- 0, -- Switchgrass, 50 cm 0, 0 0, -- 0, -- Johnsongrass, 10 cm 0, 0 0, -- 0, -- Johnsongrass, 25 cm 0, 0 0, -- 0, -- Johnsongrass, 50 cm 0, 0 0, -- 0, -- Rye grass, 10 cm 0, 0 0, -- 0, -- Rye grass, 25 cm 0, 0 0, -- 0, -- Rye grass, 50 cm 0, 0 0, -- 0, -- Other grasses, 10 cm 27.83, 3.50 18, -- 12.63, -- Other grasses, 25 cm 10.1, 2.91 4.43, -- 2.43, -- Other grasses, 50 cm 1.38, 0.91 0.9, -- 0.15, -- Woody, 1m+ 23.91, 1.23 33.65, -- 51.25, -- Habitat Vegetation category LOP OWJS RRO Dead, 5 cm 24.8, 4.64 23.24, 18.47 10.09, 2.90 Dead, 10 cm 5.04, 2.65 4.25, 3.18 1.58, 0.87 Dead, 25 cm 0.34, 0.23 0.62, 0.66 0.15, 0.15 Dead, 50 cm 0.04, 0.08 0, 0 0, 0 Forbs, 5 cm 10.84, 1.94 8.22, 4.93 18.34, 5.28 Forbs, 10 cm 6.2, 1.24 3.89, 2.32 16.25, 6.55 Forbs, 25 cm 1.53, 0.62 0.64, 0.69 8.47, 4.57 Cactus, 10 cm 0.96, 1.06 2.71, 0.11 0.1, 0.17 Cactus, 25 cm 0.77, 0.92 2.39, 0.34 0.1, 0.17 Cactus, 50 cm 0.04, 0.08 0.15, 0 0, 0 Switchgrass, 10 cm 0, 0 0, 0 0, 0 Switchgrass, 25 cm 0, 0 0, 0 0, 0 Switchgrass, 50 cm 0, 0 0, 0 0, 0 Johnsongrass, 10 cm 0, 0 0, 0 0, 0 Johnsongrass, 25 cm 0, 0 0, 0 0, 0 Johnsongrass, 50 cm 0, 0 0, 0 0, 0 Rye grass, 10 cm 0, 0 0, 0 0, 0 Rye grass, 25 cm 0, 0 0, 0 0, 0 Rye grass, 50 cm 0, 0 0, 0 0, 0 Other grasses, 10 cm 24.58, 6.07 17.67, 5.61 25.71, 14.59 Other grasses, 25 cm 9.24, 4.64 3.98, 0.64 11.45, 6.83 Other grasses, 50 cm 1.42, 0.96 0.45, 0 2.55, 2.49 Woody, 1m+ 34.01, 7.50 33.31, 1.31 4.08, 2.22 Habitat Vegetation category RW TG Dead, 5 cm 34.67, 6.60 18.5, 6.01 Dead, 10 cm 7.3, 0.47 8.84, 0.51 Dead, 25 cm 1.32, 0.16 1.14, 0.76 Dead, 50 cm 0.2, 0 0.15, 0 Forbs, 5 cm 3.67, 0.28 3.35, 0.21 Forbs, 10 cm 2.55, 0.11 2.97, 0.54 Forbs, 25 cm 0.17, 0.23 1.47, 0.54 Cactus, 10 cm 0, 0 0, 0 Cactus, 25 cm 0, 0 0, 0 Cactus, 50 cm 0, 0 0, 0 Switchgrass, 10 cm 0, 0 0.45, 0.64 Switchgrass, 25 cm 0, 0 0.53, 0.74 Switchgrass, 50 cm 0, 0 0.53, 0.74 Johnsongrass, 10 cm 0, 0 0.99, 1.40 Johnsongrass, 25 cm 0, 0 1.69, 2.39 Johnsongrass, 50 cm 0, 0 1.07, 1.51 Rye grass, 10 cm 0, 0 5.86, 2.21 Rye grass, 25 cm 0, 0 7.13, 2.23 Rye grass, 50 cm 0, 0 6.51, 0.25 Other grasses, 10 cm 4.87, 0.37 53.88, 10.08 Other grasses, 25 cm 1, 0.57 35.32, 7.16 Other grasses, 50 cm 0.2, 0 19.84, 10.83 Woody, 1m+ 46.23, 11.46 13.38, 5.13 Table 2. Summary of rodent sampling in nine habitats at Colorado Bend State Park, San Saba and Lampasas counties, Texas. Trapnights indicate sampling effort. For each rodent species and for all rodents species combined, entries include the number of individuals captured on all transects, followed by the mean relative density (number of individuals captured per 100 trapnights). Peromyscus Habitat Trapnights pectoralis Juniper/oak/persimmon mosaic 900 29, 1.59 Juniper shrub 360 25, 2.67 Juniper woods 300 14, 1.94 Oak woodland/juniper shrub 570 13, 2.83 Live oak parks 1080 9, 1.56 Rangeland/rock outcrop 750 17, 1.77 Terrace grassland 1170 23, 2.06 Riparian woodland 630 39, 7.03 Mesquite parks 180 11, 5.83 Peromyscus Sigmodon Habitat maniculatus hispidus Juniper/oak/persimmon mosaic 0, 0 1, 0.05 Juniper shrub 0, 0 0, 0 Juniper woods 0, 0 0, 0 Oak woodland/juniper shrub 0, 0 2, 0.17 Live oak parks 0, 0 1, 0.05 Rangeland/rock outcrop 0, 0 4, 0.42 Terrace grassland 1, 0.05 15, 1.1 Riparian woodland 0, 0 0, 0 Mesquite parks 0, 0 6, 4.44 Reithrodontomys Baiomys Habitat fulvescens taylori Juniper/oak/persimmon mosaic 0, 0 0, 0 Juniper shrub 0, 0 0, 0 Juniper woods 0, 0 0, 0 Oak woodland/juniper shrub 0, 0 0, 0 Live oak parks 0, 0 0, 0 Rangeland/rock outcrop 1, 0.1 0, 0 Terrace grassland 1, 0.04 1, 0.15 Riparian woodland 0, 0 0, 0 Mesquite parks 0, 0 0, 0 Chaetodipus All Habitat hispidus rodents Juniper/oak/persimmon mosaic 0, 0 30, 1.92 Juniper shrub 0, 0 25, 2.67 Juniper woods 0, 0 14, 1.94 Oak woodland/juniper shrub 0, 0 15, 3 Live oak parks 0, 0 10, 1.62 Rangeland/rock outcrop 0, 0 22, 2.29 Terrace grassland 3, 0.14 44, 3.76 Riparian woodland 0, 0 39, 7.03 Mesquite parks 0, 0 17, 10.28 Table 3. Correlation coefficients, r, for relationships of relative density of three species of rodents with vegetation categories at Colorado Bend State Park, San Saba and Lampasas counties, Texas. * indicates significance at P [less than or equal to] 0.05. Vegetation Peromyscus Sigmodon Reithrodontomys category pectoralis hispidus fulvescens Dead, 5 cm -0.26 0.10 -0.06 Dead, 10 cm -0.27 0.61 0.40 Dead, 25 cm -0.19 0.50 0.32 Dead, 50 cm < 0.01 0.35 0.26 Forbs, 5 cm -0.33 -0.16 0.13 Forbs, 10 cm -0.14 -0.11 0.25 Forbs, 25 cm -0.38 0.11 0.45 Cactus, 10 cm 0.02 -0.27 -0.48 Cactus, 25 cm 0.06 -0.22 -0.43 Cactus, 50 cm 0.29 -0.20 -0.37 Switchgrass, 10 cm -0.36 0.95 * 0.81 * Switchgrass, 25 cm -0.36 0.95 * 0.81 * Switchgrass, 50 cm -0.36 0.95 * 0.81 * Johnsongrass, 10 cm -0.36 0.95 * 0.81 * Johnsongrass, 25 cm -0.36 0.95 * 0.81 * Johnsongrass, 50 cm -0.36 0.95 * 0.81 * Rye grass, 10 cm -0.36 0.95 * 0.81 * Rye grass, 25 cm -0.36 0.95 * 0.81 * Rye grass, 50 cm -0.36 0.95 * 0.81 * Other grasses, 10 cm -0.58 0.87 * 0.81 * Other grasses, 25 cm -0.53 0.94 * 0.88 * Other grasses, 50 cm -0.40 0.97 * 0.86 * Woody, 1m+ 0.50 -0.60 -0.77 *
This project was funded by Texas Parks and Wildlife, contract no. 89758, and administered by Keith Blair and David Riskind. We thank Tracy Carter, Cathy Early, Chris Filstrup, Shannon Hill, Michael Mellon, Matt Schaub and Shannon Taylor for their invaluable help in the field. Cory Evans, and the other park employees at Colorado Bend State Park, kindly provided assistance and lodging during our fieldwork. We also thank Dr. W. K. Hartberg, chairman of Baylor Biology Department, for the use of field vehicles and other equipment, and Prof. W. C. Holmes for his identification of grasses.
Baccus, J. T. & J. K Horton. 1984. Habitat utilization by Peromyscus pectoralis in central Texas. Pp. 7-26, in Festschrift for Walter W. Dalquest in honor of his sixty-sixth birthday (N. V. Homer, ed.). Midwestern State University Press, Wichita Falls, Texas, 163 pp.
Bailey, V. 1905. Biological survey of Texas. N. Am. Fauna, vol. 25. Washington, D.C.: Dept. of Agriculture, Bureau of Biological Survey, 222 pp.
Blair, W. F. 1950. The biotic provinces of Texas. Texas J. Sci., 2(1):93-117.
Daubenmire, R. 1959. A canopy coverage method of vegetational analysis. Northwest Sci., 33:43-64.
Davis, W. B. & D. J. Schmidly. 1994. The mammals of Texas. Texas Parks and Wildlife Department, Austin, Texas, x + 338 pp.
Etheredge, D. R., M. D. Engstrom & R. C. Stone, Jr. 1989. Habitat discrimination between sympatric populations of Peromyscus attwateri and Peromyscus pectoralis in west-central Texas. Journal of Mammalogy, 70:300-307.
Gehlbach, F. R. 1991. The east-west transition zone of terrestrial vertebrates in central Texas: a biogeographical analysis. Texas J. Sci., 43(4):415-427.
Goetze, J. W. 1998. The mammals of the Edwards Plateau, Texas. Spec. Pub., Mus. of Texas Tech Univ., 41:1-263.
Gould, F. W. 1975. Texas plants: a checklist and ecological summary. Texas Agricultural Experiment Station, MP-585/Revised, 121 pp.
Hanchey, M. F. & K. T. Wilkins. 1998. Habitat associations of the small-mammal community in the Grand Prairie of north-central Texas. Texas J. Sci., 50(2): 107-122.
Pimentel, R.A. 1993. Bio[SIGMA]tat III Reference Manual. SigmaSoft, 116 pp.
SAS Institute. 1985. SAS user's guide: basics, 5th ed. SAS Institute Inc., Cary, N.C., 1,292 pp.
Schmidly, D. J. 2002. Texas natural history: a century of change. Texas Tech University Press, Lubbock, xiv + 534 pp.
U.S.D.A. 1982. Soil survey of San Saba County, Texas. United States Department of Agriculture, Soil Conservation Service and Texas Agricultural Experiment Station, 163 pp. + maps.
Wilkins, K. T. 1995. The rodent community and associated vegetation in a tallgrass blackland prairie in Texas. Texas J. Sci., 47(4):243-262.
KTW at: email@example.com
Jeffrey A. Scales and Kenneth T. Wilkins, Department of Biology and Graduate School Baylor University Waco, Texas 76798-7388
|Printer friendly Cite/link Email Feedback|
|Author:||Scales, Jeffrey A.; Wilkins, Kenneth T.|
|Publication:||The Texas Journal of Science|
|Date:||Aug 1, 2003|
|Previous Article:||Reproduction in the speckled racer, Drymobius margaritiferus (Serpentes: Colubridae) from Mexico and Central America.|
|Next Article:||A graphical representation for inverse functions.|