Effects of cattle grazing on salt desert rodent communities.
Cattle grazing can have profound effects on physiognomy and composition of desert plant communities. Long-term grazing has been shown to decrease the abundance of perennial grasses and forbs and increase the amount of annual grasses and weeds in deserts (Rice and Westoby, 1978; Brotherson and Brotherson, 1981; Hanley and Page, 1981; Medin and Clary, 1990). Cattle grazing also decreases the amount of litter (Milchunas et al., 1992), and moderate to intense grazing increases soil bulk density (Van Harren, 1983) and decreases soil aggregate stability (Warren et al., 1985) in desert regions.
Cattle grazing also can impact small mammal communities. Hanley and Page (1981) found that grazing decreased rodent species diversity in arid environments, probably due to a decline in plant species diversity that resulted from the grazing treatment. Rosenzweig and Winakur (1969) also found a negative correlation between grazing intensity, and rodent species diversity in arid regions. However, they attributed this to changes in structural aspects of vegetation rather than plant species diversity.
General knowledge about the natural history of desert rodents can be used to predict potential effects of grazing on their communities. For example, bipedal kangaroo rats (Dipodomys spp.) and kangaroo mice (Microdipodops spp.) are often associated with open (Rosenzweig, 1977; Price and Brown, 1983), and disturbed (Mellink, 1985; Stangl et al., 1992) areas. In contrast, pocket mice (Perognathus spp. and Chaetodipus spp.) typically forage under perennial vegetation (Price and Brown, 1983; Reichman and Price, 1993), possibly because they are less adept at securing seed resources and/or at evading predators in open spaces (Price and Brown, 1983; Longland and Price, 1991). Thus, one might predict, based on rodent natural histories, that cattle grazing may favor bipedal heteromyids more than quadrupedal species because of grazing-induced changes in vegetative structure.
Heteromyids and other rodent species in North America respond to variability in seed production (Brown, 1973, 1975; French et al., 1974). In general, grazing-induced changes in densities of grasses, forbs and exotic weeds in arid environments may affect foraging strategies of rodents, but the specific response depends on the degree of impact and life-histories of the plants and rodents present. Heteromyid rodents can harvest the vast majority of the seed production of certain plant species (Chew and Chew, 1970; Soholt, 1973) and may be important in seed dispersal and seedling establishment of certain native plants by caching seeds (Vander Wall, 1990; Longland, 1995). In addition, selective feeding by rodents on competitively superior plants and their seeds can be important in maintaining a heterogeneous plant community (Brown and Heske, 1990). Major alterations in abundance of desert rodents could thus have an undesirable cascading effect on the plant community. Given this potential keystone role of rodents in North American deserts, diverse rodent species assemblages could provide important indicators of rangeland ecosystem integrity.
We tested whether different levels of grazing and associated effects of cattle activity influenced the composition and activity patterns of rodent species in a salt desert community. We first examined potential differences in relative abundance of rodents between lightly and heavily grazed portions of grazing-intensity gradients associated with surface wells. Because wells are often the only sources of water for cattle in salt desert environments, cattle typically do not move more than a few kilometers away from wells. Therefore, vegetation close to a well is usually heavily impacted by grazing, trampling and increased nitrogen input from feces, and vegetation 4-5 km away is lightly impacted (J. A. Young, pets. comm.). For convenience, we discuss all cattle impacts (trampling, nitrogen input, etc.) in terms of "grazing" effects rather than restricting this terminology to just effects directly resulting from herbivory. In addition to testing for population and community-level effects, we tested for two potential effects of cattle on activity patterns of certain species. First, we tested for differences in sizes of rodent home ranges in lightly and heavily grazed areas. Second, we assessed microhabitat selection by rodents, and analyzed vegetation to determine whether cattle grazing affected microhabitat attributes. For all the above comparisons, we tested the null hypothesis that heavily grazed areas did not differ from lightly grazed areas.
Although some studies have assessed cattle influences on abundances of rodents (e.g., Bock et al., 1984; Medin and Clary, 1990), none has determined how grazing may affect microhabitats used by rodents. In addition, few published studies have used multiple sites to investigate grazing effects. A single fence line comparison between a nongrazed exclosure and an adjacent grazed area is typical (Cottam and Evans, 1945; Gardner, 1950; Brady et al., 1989). In this study we compared lightly and heavily grazed portions of three grazing gradients. True replication may not be possible with such field studies (Hurlbert, 1984), in our case because of between-site differences in stocking rates, times of grazing, and soils. However, any within-site differences between heavily and lightly grazed areas of our gradients that are concordant among the different gradients (i.e., sites) constitute strong evidence of grazing effects.
STUDY AREA AND METHODS
Study site. - The study was conducted from 1993 to 1995 in two locations in Nightingale Flat Valley, 35 km N and 30 km E of Fernley, Churchill County, Nev. (elev. 1350 m) and one location in Little Valley, 18 km N of Wadsworth, Washoe County, Nev. (elev. 1220 m).
Study sites were located in salt desert vegetation dominated by shadscale (Atriplex confertifolia) and Bailey's greasewood (Sarcobatus vermiculatus baileyi), with winterfat (Kraschen-innikovia lanata) in the shrub layer. The herbaceous understory was sparse with coarser textured soils supporting Indian ricegrass (Achnatherum hymenoides). Precipitation varied in the region over the course of the study. From 1 November 1993 - April 1994 precipitation at the nearest weather station was 0.43 cm; October 1994 - April 1995 precipitation was 1.93 cm (National Weather Service precipitation data, 1993-1995).
We identified at least one grazing-intensity gradient defined by a well at each of these sites (Telephone well gradient and Geothermal well gradient in Nightingale Flat Valley and Pyramid well gradient in Little Valley). Each well occurred on a separate grazing allotment, and all three allotments have been traditionally winter grazed by 50 to 150 cattle. All allotments were grazed at least once during or immediately preceding the 17-mo study. We set up a 10-by-10 station trapping grid with 15 m trap spacing in the heavily and lightly grazed portion of each gradient for a total of six grids (3 gradients with 2 grids per gradient). Heavily-grazed grids were located within 300 m of each well, but outside the immediate vicinity of the well that was nearly denuded of vegetation by cattle. Lightly-grazed grids were 3-5 km from wells. We walked a series of 1-km by 1-m linear transects to select gradients where there were differences of approximately 5-10 fold in densities of cattle feces between the well and the area chosen as lightly grazed. In addition, lightly- and heavily-grazed portions of each gradient were matched for vegetation associations and soil type, including the presence of rocky substrates.
In June and July 1994 we sampled vegetation at 20 randomly selected trap stations on each 100-station grid at the Geothermal and Telephone gradients. These stations served as the centers for 100 [m.sup.2] (5.64-m radius) circular plots within which we sampled cover, frequency and density of vegetation. Using line intercepts, we estimated percent vegetation cover and percent bare ground along the N-S and E-W diameters of the circular plots to the nearest percent. Frequencies of occurrence of perennial grass, cheatgrass (Bromus tectorum), cryptogam crust, and native forbs also were measured by systematically placing 10, 1-[m.sup.2] frames along the N-S and E-W axes of each circular plot. Sampling was conducted early in the summer to accurately assess annual plant frequencies. We determined shrub densities by counting the number of shrubs in each circular plot. In May-July 1995 similar analyses to those conducted at the Geothermal and Telephone gradients in 1994 were performed at the Pyramid gradient. In addition, vegetation frequency and shrub density were resampled on all grids in 1995, and shrub densities were recorded separately for the two most common shrubs (Bailey's greasewood and Atriplex spp.).
We used one-way analyses of variance (ANOVAs) to compare vegetation characteristics of the lightly- and heavily-grazed areas within each grazing gradient (GLM procedure; SAS Institute, Inc., 1987). Two-way ANOVAs were performed on combined vegetation data for all gradients, with a site variable added to these analyses. Because different vegetation measures may not be independent of one another (i.e., percent cover bare ground is dependent on percent cover shrubs, grasses, etc.), we used multivariate analysis of variance (MANOVA) to test whether the treatment effect (grazing) was independent of correlations among the vegetation categories (SAS Institute, Inc., 1987).
Trapping studies. - To determine overall rodent numbers and relative species abundances in lightly- and heavily-grazed areas, we live-trapped all three gradients in 1994 and 1995. The Pyramid gradient was trapped 3 consecutive nights (hereafter referred to as a session), once per month, August-October 1994, and June-October 1995. The Telephone and Geothermal gradients were trapped one session per month, July-October 1994, and May-September 1995. On trap evenings, one Sherman live trap (7.6 x 8.9 x 22.9 cm) baited with wild bird seed mixture was set at each trap station. The next morning we recorded species, trap location and sex for each captured rodent; all rodents were marked with a uniquely numbered eartag upon first capture. Although we concentrated on nocturnal species, traps were open for 1-3 h of daylight after baiting and before data collection. We caught reasonable numbers of the only diurnal species present at our sites (Ammospermophilus leucurus), and used these data as well as those for nocturnal species in our analyses.
Repeated measures ANOVA (SAS Institute, Inc., 1987) with treatment nested within site was used to test for differences in numbers of animals per species between the lightly- and heavily-grazed areas for the four most commonly captured rodent species. Class variables were rodent species, grazing treatment, and site (i.e., gradient), with session as a repeated factor. There were two sessions common to all three sites in 1994 (August, September) and four common sessions in 1995 (June-September), and these data were all included in the analysis. The species, treatment and species x treatment interaction terms in these analyses were tested over the species x treatment interaction nested within site as an error term. We used the mean number of individual animals captured per night per session as the dependent variable, rather than the number of total captures including recaptures to avoid skewing data towards those individuals that were recaptured most frequently. We used Tukey's Studentized Range test to separate differences among species/treatment combinations in cases of significant species-by-treatment interactions.
We also calculated maximum movement distances as an index of home-range size for animals captured three or more times. Differences in maximum movement distances of the more common species (Dipodomys merriami, D. microps, Perognathus longimembris and Microdipodops megacephalus) between the lightly- and heavily-grazed areas were analyzed separately for each grazing gradient with Student's t-tests. To reduce the chance of type I errors, we applied a sequential Bonferroni adjustment to these tests (Rice, 1989).
Studies of microhabitat use. - To determine microhabitat affinities of various species, we conducted night-time tracking studies using techniques described in detail by Longland and Clements (1995). After sundown, we placed trays lined with sandpaper covered with pigmented fluorescent powder near the trapping grids in the lightly- and heavily-grazed portions of the grazing gradients. In the middle of each tray was a petri dish with 40 g of millet seeds that also were covered with fluorescent powder. Once a rodent started to harvest the seeds, we noted the species of animal, and 30 min later we followed the animal's footprints in the substrate using a UV lantern. We marked each trail with pinflags, and used a unique color of flag to mark stopping points. These points were apparent due to scattered dye on vegetation and the soil surface, and indicated that the rodent stopped to cache seeds, dig, or harvest more seeds.
Data were collected on 273 m of rodent trails representing 19 individuals and four species (Dipodomys merriami, D. microps, Perognathus longimembris and P. parvus). Stopping points served as centers of 1-m radius circles, within which we recorded cover of bare ground, shrubs, perennial grasses and exotic weeds using the line-intercept method described earlier. Data on vegetation cover from trail experiments were compared to cover data for the same areas in the general vegetation analyses. We found it difficult to distinguish D. merriami and D. microps during these trials unless we had the animal in hand, which was not always possible. Therefore, we tested for heteromyid rodent affinities from trail data at the level of genera by combining data for the two Dipodomys species and the two Perognathus species. While these two heteromyid genera tend to exhibit the typical biped-quadruped microhabitat differences, species within each genus use microhabitat features similarly (Price and Brown, 1983). Student's t-tests were used to compare each measure (i.e., percent cover bare ground, percent cover shrubs etc.) of the rodent stopping points to the same measures in the general vegetation analysis at that site to determine whether these different rodent genera were using various microhabitat features in frequencies that differed from those available in the surrounding area. We also applied a sequential Bonferroni adjustment to these analyses (Rice, 1989). For purposes of discussion, we report microhabitat use of Dipodomys and Perognathus in terms of percent (%) use of microhabitat features relative to their availability. For example, if the combined average of pocket mice stopping points intersected shrub habitat 40% of the time, but shrubs only comprised 30% of the cover of the general surrounding area, then we report that pocket mice used 133% of the expected amount of shrub habitat.
Vegetation analyses. - In six of seven vegetation analyses there were significant treatment effects that were not masked by the potential interrelatedness of the various vegetative categories (Wilks' lambda, P [less than] 0.05 for these six analyses). Therefore, we used univariate statistics to describe vegetation differences between the lightly and heavily grazed areas.
In 1994, 9 of 14 vegetation categories differed significantly between lightly and heavily grazed areas (Table 1). There was greater shrub cover in lightly grazed areas ([F.sub.(1,57)] = 23.57, P [less than] 0.001) due mainly to the Telephone gradient. Frequencies of forbs ([F.sub.(1,57)] = 11.20, P = 0.002) and cryptogam crusts ([F.sub.(1,57)] = 126.90, P [less than] 0.001) were greater in lightly than heavily grazed areas. Cover of exotic annuals was greater in the heavily grazed areas ([F.sub.(1,57)] = 4.93, P = 0.03), as was the frequency of Indian ricegrass ([F.sub.(1,57)] = 6.56, P = 0.013), and the cover ([F.sub.(1,57)] = 12.05, P = 0.001), frequency ([F.sub.(1,57)] = 15.46, P [less than] 0.001) and density of winterfat ([F.sub.(1,57)] = 31.03, P [less than] 0.001).
Results of 1995 vegetation surveys were similar to those of 1994. In 1995, five of seven vegetation categories measured were significantly different between lightly and heavily grazed areas. The frequencies of cryptogam crust ([F.sub.(2,95)] = 75.57, P [less than] 0.001) were higher in lightly grazed areas, and frequencies of ricegrass ([F.sub.(2,95)] = 15.63, P [less than] 0.001), exotic annuals ([F.sub.(2,95)] = 24.39, P [less than] 0.001) and winterfat (F.sub.(2,95)] = 3.87, P = 0.024) were higher in heavily grazed areas (Table 1). Densities of Bailey's greasewood varied in the lightly and heavily grazed areas, but saltbush (all Atriplex spp.) was more abundant ([F.sub.(2,95)] = 23.12, P [less than] 0.001) in lightly grazed areas.
Trapping studies. - Eleven species of rodents were captured, including eight species in the Family Heteromyidae: Dipodomys deserti, D. merriami, D. microps, D. ordii, Microdipodops megacephalus, Chaetodipus formosus, Perognathus longimembris and P parvus. The non-heteromyid species included Ammospermophilus leucurus, Onychomys torridus and Peromyscus maniculatus.
Only the four species most frequently captured at all three sites (Ammospermophilus leucurus, Dipodomys merriami, D. microps, Perognathus longimembris) were used in the analyses. A species-by-treatment interaction was found for numbers of rodents captured in both 1994 and 1995 (Table 2). This interaction can largely be explained by consistently greater numbers of D. merriami (P [less than] 0.05, Tukey's test) in heavily grazed areas and of P. longimembris (P [less than] 0.05, Tukey's test) in the lightly grazed areas [ILLUSTRATION FOR FIGURE 1 OMITTED].
Due to small samples of maximum-movement-distance (MMD) data collected at the Telephone and Pyramid gradients, comparisons between grazing treatments and among species [TABULAR DATA FOR TABLE 1 OMITTED] could only be made at the Geothermal gradient, for which we had MMD data for four species in 1994 and three species in 1995. MMDs were 1.5-2.0 times greater in the heavily grazed area for all four species in 1994 (Wilcoxon signed ranks test: T+ = 10, N = 4, P = 0.062, which is the lowest P value achievable with this small sample size; Siegel, 1956), but there was no such trend in 1995.
Microhabitat use studies. - t-tests were only possible for Dipodomys data from Pyramid-heavy [TABULAR DATA FOR TABLE 2 OMITTED] and Geothermal-light grids and for Perognathus data from Telephone-heavy and Geothermal-light grids, because these areas had data for at least four individuals that could be tested against cover data from vegetation analyses. The amount of bare ground at kangaroo rat stopping points was nearly identical to that expected based on general vegetation (101% of expected), but stopping points had only 38% of expected shrub cover (shrub use at Pyramid-heavy: t = 4.05, df = 29, P [less than] 0.001; Geothermal-light: t = 2.34, df = 23, P = [0.026.sup.2]). Pocket mice used stopping points with 63% of the expected amount of bare ground (Geothermal-light: t = 3.31, df = 22, P = 0.004; Telephone-heavy: t = 1.80, df = 23, P = 0.079) and only 13% of expected Indian ricegrass cover (Geothermal-light: t = 4.22, df = 22, P [less than] 0.001; Telephone-heavy: t = 2.09, df = 23, P = [0.042.sup.2]). Pocket mice used areas with 134% of expected shrub cover, and although this did not differ significantly from expected (Geothermal-light: t = 1.52, df = 22, P = 0.16; Telephone-heavy: t = 0.34, df = 23, P = 0.71), our findings of significant under-use of bare ground by pocket mice and of shrubs by kangaroo rats indicate that pocket mice used vegetated microhabitats while kangaroo rats frequented open areas.
We found significantly more Dipodomys merriami in heavily grazed areas in 1994 and more Perognathus longimembris in lightly grazed areas in 1995. Desert rodent species often differ in affinities for different soil and vegetation features. Specialization on specific habitat features can cause dramatic changes in density of a particular rodent species or even replacement of one species by another over relatively small spatial scales (Price and Brown, 1983). Due to the absence of any noticeable soil or topographic differences between lightly- and heavily-grazed areas and the similarity in extrinsic environmental effects over the small distances separating our paired trapping grids, rodent populations between lightly- and heavily-grazed habitats are most likely the result of vegetation differences associated with grazing intensity. There were more shrubs in the lightly grazed areas; this may explain the greater abundance of P longimembris in these areas, as Perognathus spp. are known to prefer habitats with high shrub cover (Price and Brown, 1983; Reichman and Price, 1993). Furthermore, it has been suggested that when there is sufficient cover, Perognathus spp. may depress densities of D. merriami perhaps due to greater efficiency of procuring seeds (Rosenzweig and Winakur, 1969; Stamp and Ohmart, 1977).
Even though Dipodomys merriami was more abundant near wells, D. microps usually occurred in greater numbers in lightly grazed areas (Fig. 1). Reduced abundance of D. microps in a grazed area may be due to reduction in saltbush abundance in heavily-grazed areas since saltbush foliage is an important dietary item for D. microps (Kenagy, 1973). Cattle will feed on saltbush, especially saltbush seedlings, if other desirable types of forage are not available (Range Plant Handbook, 1937).
Maximum-movement-distance (MMD) results might also be explained by different grazing intensities. At the Geothermal gradient in 1994, all species had higher MMDs in the heavily-grazed area. This may be the result of individuals moving longer distances to search for resources at this site. King (1968) and Scheibe (1984) found that desert rodents had enlarged home ranges when resources were lacking. While frequency of Indian ricegrass, a highly preferred seed resource (McAdoo et al, 1983; Kelrick et al., 1986), was greater at the heavily-grazed geothermal grid, cover of ricegrass was similar at both grids, because the lightly grazed grid had larger ricegrass clumps. The findings that Indian ricegrass plants were more frequent, but provided similar cover, in the heavily-grazed relative to lightly-grazed areas support our casual observation that clumps of Indian ricegrass were more robust in the lightly-grazed areas. If seed production of Indian ricegrass increases exponentially with the size of grass clumps, then the larger-clumped, lightly-grazed populations would yield substantially more seeds than smaller-clumped, heavily-grazed populations, even though ricegrass cover was statistically similar in these populations. This is potentially important because Indian ricegrass seeds are known to be a preferred resource for desert rodents (McAdoo et al., 1983; Kelrick et al., 1986). We tested data from another study at a nearby field site for an effect of Indian ricegrass clump size on seed numbers and mass using both linear and exponential models with stepwise regression (SAS REG Procedure, SAS Institute Inc., 1987). For all models tested (using either seed mass or seed numbers and either a linear or exponential model), the stepwise procedure only included the interaction between grass clump circumference and height as a significant predictor of seed production, and omitted main effect terms. For both dependent variables, the exponential model was the superior predictor of seed production (seed mass: [F.sub.(1,29)] = 209.52, P [less than] 0.001, [R.sup.2] = 0.88; numbers: [F.sub.(1,29)] = 251.21, P [less than] 0.001, [R.sup.2] = 0.90) compared with the linear model (seed mass: [F.sub.(1,29)] = 88.29, P [less than] 0.001, [R.sup.2] = 0.76; numbers: [F.sub.(1,29)] = 158.12, P [less than] 0.001, [R.sup.2] = 0.85). This shows that larger clumps of Indian ricegrass produce seeds at a rate proportionally greater than that of smaller clumps. If rodents at the Geothermal-heavy grid had to move greater distances than those at Geothermal-light to harvest similar amounts of Indian ricegrass seeds, this could explain greater MMDs in the former area.
The trail studies provided the most insight into effects of cattle grazing on the various rodent species and their microhabitats. Our trail data agree with previous investigations reporting relatively greater affinities of Dipodomys spp. for open habitats and of Perognathus spp. for vegetated microhabitats (Price and Brown, 1983; Reichman and Price, 1993). At our sites, shrubs were generally more abundant in lightly grazed areas. In addition, Dipodomys merriami, the most abundant kangaroo rat at the study sites, was more numerous in the heavily grazed treatments, whereas P. longimembris, the most abundant pocket mouse at all sites, was more numerous in the lightly grazed treatments. Our results suggest that cattle, by foraging either on shrub seedlings or fruit, can reduce overall shrub density, thus favoring conditions for D. merriami but not Perognathus spp. This is consistent with other studies that have shown that grazing in deserts can lead to decreased shrub densities (Brotherson and Brotherson, 1981; Bock et al., 1984). Furthermore, other studies also have found Perognathus spp. to be more abundant in ungrazed habitats (Hanley and Page, 1981; Bock et al., 1984) and Dipodomys spp. to be more abundant in grazed habitats (Reynolds, 1958; Bock et al., 1984).
During the study one species (Perognathus longimembris) dramatically increased in numbers on all grids, whether heavily- or lightly-grazed. Increased precipitation during the winter preceding the 1995 trap season was probably the factor leading to the pronounced abundance of P longimembris during this season. Comparison of these years with prior winters suggest that the winter of 1993-1994 experienced less than average precipitation and the winter of 1994-1995 greater than average precipitation. Beatley (1969, 1976) showed a correlation between autumn/winter precipitation and productivity of winter annuals, which in turn leads to increased desert rodent reproduction. Perognathus longimembris reacted positively to increased precipitation in the winter of 1994-1995, but the abundance of this species was still significantly reduced by heavy grazing regardless of the population increase.
Although there were no major differences in rodent species richness in lightly- vs. heavily-grazed areas, grazing did result in a shift in species composition through its impacts on resources that are important to those species. Rodent communities in nature are intrinsically dynamic, and some of the dynamics noted during the relatively short duration of our study differed systematically between heavily- and lightly-grazed areas, indicating that these animals are sensitive to livestock grazing. It is thus possible that long-term mismanagement will have severe impacts on some species that extend well beyond the relatively subtle grazing effects we found. Ranchers should manage stocking rates so that bunchgrasses can set adequate seed to sustain rodent populations, and so that cattle have adequate forage to prevent feeding on less preferred, sensitive plants such as saltbush.
Because our study used no "true" ungrazed controls, we believe that our results are conservative in their assessment of grazing effects on rodent communities. The fact that we could not find an adequate number of suitable, ungrazed reference sites for this study implies that the characterization of livestock grazing on arid western rangelands as a "uncontrolled experiment" (Noss, 1994) is not simply an alarmist slogan.
Acknowledgments. - Helpful suggestions for this manuscript were provided by J. L. Rachlow, S. B. Vander Wall, J. M. Reed, S. H. Jenkins and J. A. Young. We also have great appreciation for the many who provided field assistance for this project D. Palmquist was instrumental in this project for her help with statistical analyses. We thank the American Museum of Natural History for financial support for this project. This material is based upon work supported by the Cooperative State Research Service, U.S. Department of Agriculture, under Agreement No. 93-37101-8995. This paper is a contribution of the U.S. Department of Agriculture, Agricultural Research Service, Conservation Biology. of Rangelands Unit, Reno, Nevada.
2 This value is not significant after the Bonferroni adjustment.
BEATLEY, J. C. 1969. Dependence of desert rodents on winter annuals and precipitation. Ecology, 50: 721-724.
-----. 1976. Rainfall and fluctuating plant populations in relation to distribution and numbers of desert rodents in southern Nevada. Oecologia, 24:21-42.
BOCK, C. E, J. H. BOCK, W. R. KENNEY AND V. M. HAWTHORNE. 1984. Responses of birds, rodents and vegetation to livestock exclosure in a semidesert grassland. J. Range Manage., 37:239-244.
BRADY, W. W, M. R. STROMBERG, E. F. ALDON, C. D. BONHAM AND S. H. HENRY. 1989. Response of a semidesert grassland to 16 years of rest from grazing. J. Range Manage., 42:284-288.
BROTHERSON, J. D. AND W. W. BROTHERSON. 1981. Grazing impacts on the sagebrush communities of central Utah. Great Basin Nat., 41:335-340.
BROWN, J. H. 1973, Species diversity of seed-eating desert rodents in sand dune habitats. Ecology, 54: 775-787.
-----. 1975. Geographical ecology of desert rodents, p. 315-341. In: M. L. Cody and J. M. Diamond (eds.). Ecology and evolution of communities. Harvard University Press, Cambridge, Mass.
----- AND E. J. HESKE. 1990. Control of a desert-grassland transition by a keystone rodent guild. Science, 250:1705-1707.
CHEW, R. M. AND A. E. CHEW. 1970. Energy relationships of mammals of a desert shrub (Larrea tridentata) community. Ecol. Monogr., 40:1-21.
COTTAM, W. P. AND F. R. EVANS. 1945. A comparative study of the vegetation of grazed and ungrazed canyons of the Wasatch Range, Utah. Ecology, 26:171-181.
FRENCH, N. R, B. G. MAZA, H. O HILL, A. P. SCHWANDEN AND H. W. KAAZ. 1974. A population study of irradiated desert rodents. Ecol. Monogr., 44:45-72.
GARDNER, J. L. 1950. Effects of thirty years of protection from grazing in desert grassland. Ecology, 31: 44-50.
HANLEY, T. H. AND J. L. PAGE. 1981. Differential effects of livestock use on habitat structure and rodent populations. Calif Fish Game, 68:160-173.
HURLBERT, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr., 54:187-211.
KELRICK, M. I., J. A. MACMAHAN, R. R. PARMENTER AND D. V. SISSON. 1986. Native seed preferences of shrub-steppe rodents, birds and ants: the relationship of seed attributes and seed use. Oecologia, 68:327-337.
KENAGY, G. I. 1973. The periodicity of daily activity and its seasonal changes in free-ranging and captive kangaroo rats, Oecologia, 24:105-140.
KING, J. A. 1968. Biology of Peromyscus. Spec. Publ. Am. Soc. of Mammalogists., 2:1-593.
LONGLAND, W. S. 1995. Desert rodents in disturbed shrub communities and their effects on plant recruitment. p. 209-215. In: B. A. Roundy, E. D. McArthur, J. S. Haley and D. K. Mann (eds.). Proceedings of the Symposium on wildland shrubs and arid land restoration. U.S. For. Ser. Gen. Tech. Rep, INT-GTR-315.
----- AND C. CLEMENTS. 1995. Use of fluorescent pigments in studies of seed caching by rodents. J. Mammal, 76:1260-1266.
----- AND Mi V. PRICE. 1991. Direct observations of owls and heteromyid rodents: can predation risk explain microhabitat use? Ecology, 72:2261-2273.
MCADOO, J. K., C. C. EVANs, B. A ROUNDY, J. A. YOUNG AND R. A. EVANS. 1983. Influence of Heteromyid rodents on Oyrzopsis hymenoides germination. J. Range Manage., 36:61-64.
MEDIN, D. AND W. CLARY. 1990. Small mammal populations in a grazed and ungrazed riparian habitat in Nevada. U.S. Int. Res. Pub. Res. Pap. INT-413., U.S. For. Ser. Ogden, Utah.
MELLINK, E. 1985. Agricultural disturbance and rodents: three farming systems in the Sonoran desert. J. Avid Env., 8:207-222.
MILCHUNAS, D. G., W. K. LAUENROTH AND P. L. CHAPMAN. 1992. Plant competition, abiotic, and long and short term effects of large herbivores. Oecologia, 92:520-531.
NOSS, R. F. 1994. Cows and conservation biology. Conserv. Biol., 8:613-616.
PRICE, M.V. AND J. H. BROWN. 1983. Patterns of morphology and resource use in North American desert rodent communities. Great Basin Nat. Mem., 7:117-134.
RANGE PLANT HANDBOOK. 1937. U.S. For. Ser. U.S. Government Printing Office, Washington, D.C.
REICHMAN, O. J. AND M. V. PRICE. 1993. Ecological aspects of heteromyid foraging, p. 539-568. In: H. H. Genoways and J. H. Brown (eds.). Biology of heteromyid rodents. Spec. Publ. No. 10. The American Society of Mammalogists.
REYNOLDS, H. G. 1958. The ecology of the Merriam kangaroo rat on the grazing lands of southern Arizona. Ecol. Monogr., 28:111-127.
RICE, B. AND M. WESTOBY. 1978. Vegetative responses of some Great Basin shrub communities protected against domestic stock. J. Range Manage., 31:28-34.
RICE, W. R. 1989. Analyzing tables of statistical tests. Evolution, 43:223-225.
ROSENZWEIG, M. L. 1977. Coexistence and diversity in heteromyid rodents, p. 89-99. In: B. Stoehouse and C. Perrins (eds.). Evolutionary ecology. University Park Press. Baltimore, Md.
----- AND J. WINAKUR. 1969. Population ecology of desert rodent communities: habitats and environmental complexity. Ecology, 50:558-572.
SAS INSTITUTE, INC. 1987. SAS/STAT guide for personal computers, version 6, 4th ed. SAS Institute, Inc. Cary, N.C. 1028 p.
SCHEIBE, J. S. 1984. Sexual differences in the home ranges of Peromyscus truei and Dipodomys panamantinus. Southwest. Nat., 29:7-13.
SIEGEL, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill, New York, N.Y. 312 p.
SOHOLT, L. F. 1973. Consumption of primary production by' a population of kangaroo rats (Dipodomys merriami) in the Mojave Desert. Ecol. Monogr., 43:357-376.
STAMP, N. E. AND R. D. OHMART. 1977. Resource utilization by desert rodents in the lower Sonoran desert. Ecology, 59:700-707.
STANGL, F. B., T. S. SCHAFER, J. R. GOETZE AND W. PINCHAK. 1992. Opportunistic use of modified and disturbed habitat by the Texas kangaroo rat. Tex. J. Science, 44:25-35.
VANDER WALL, S. B. 1990. Food hoarding in animals. University of Chicago Press, Chicago, Ill. 445 p.
VAN HARREN, B. P. 1983. Soil bulk density as influenced by grazing intensity. J. Range Manage., 36: 586-591.
WARREN, S. D., M. B. MERRIL, W. H. BLACKBURN AND N. E. GARZA. 1985. Soil response to trampling under intensive rotation grazing. Soil Sci. Soc. Am. J., 50:1336-1341.
|Printer friendly Cite/link Email Feedback|
|Author:||Jones, Allison L.; Longland, William S.|
|Publication:||The American Midland Naturalist|
|Date:||Jan 1, 1999|
|Previous Article:||The foraging ecology of the gray rat snake (Elaphe obsoleta spiloides) - visual stimuli facilitate location of arboreal prey.|
|Next Article:||Growth rates and intraspecific variation in body weights of raccoons (Procyon lotor) in southern Texas.|