Great Basin Naturalist

Volume 49 Number 3 Article 14

7-31-1989

Influence of experimental habitat manipulations on a desert rodent population in southern Utah

Jiping Zou Brigham Young University

Jerran T. Flinders Brigham Young University

Hal L. Black Brigham Young University

Steven G. Whisenant Texas A&M University

Follow this and additional works at: https://scholarsarchive.byu.edu/gbn

Recommended Citation Zou, Jiping; Flinders, Jerran T.; Black, Hal L.; and Whisenant, Steven G. (1989) "Influence of experimental habitat manipulations on a desert rodent population in southern Utah," Great Basin Naturalist: Vol. 49 : No. 3 , Article 14. Available at: https://scholarsarchive.byu.edu/gbn/vol49/iss3/14

This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Great Basin Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. INFLUENCE OF EXPERIMENTAL HABITAT MANIPULATIONS ON A DESERT RODENT POPULATION IN SOUTHERN UTAH

1 2 4 JipingZou , JerranT. Flinders . Hal L. Black', and Steven G. Whisenant

Abstract. —This paper addresses how habitat manipulations in a black sagebrush (Artemisia nova) -dominated area, John s Valley of southern Utah, affected resident desert rodent populations. Rodents studied included the deer mouse {Peromyscus maniculatus), Great Basin pocket mouse (Perognatkus partus), sagebrush vole (Lagurus curtatus), Ord's kangaroo rat (Dipodomys ordii), and least chipmunk (Eutamias minimus). The experimental design involved analyses of treatment and control (nontreatment) plots rather than pre- and posttreatment of all plots. Habitat manipulations emphasized cutting of shrubs (rotobeating), treatment of with a herbicide (2,4-D), and reseeding with a mixture of grasses, forbs, and shrubs. Posttreatment trapping indicated the deer mouse was the most abundant rodent in treatment and control plots. Data indicate the prescribed habitat treatments had no significant negative affects on the deer mouse demes on the control or treatment plots. Habitat treatments may have negatively impacted recruitment in pocket mice. Least chipmunks were not captured in plots treated by rotobeating. Our habitat manipulations may have contributed to interspecific competition in this rodent community through the reduction of both food and cover.

Many ecological factors influence the dy- supplies and essential microhabitat (Taylor namics of a cold desert rodent population. 1963, Weckey 1963, Holbrook 1979). Alterna- When aspects of the vegetational habitat are tive conclusions resulting from Parmenter altered for management purposes, subse- and MacMahon's (1983) habitat manipulations quent responses can be expected in rodent within a shrub steppe ecosystem suggest that populations that reflect changes in inter- shrub architecture and other shrub-related specific relationships as well as changes in resources are unimportant to some rodents density, home range, reproduction, disper- (e.g., deer mouse, Peromyscus maniculatus, sal, recruitment, body size, and food habits. It Great Basin pocket mouse, Perognatkus is difficult to relate measured responses to parvus; and northern grasshopper mouse, specific changes of components in the habitat. Onychomys leucogaster). Others have shown Reason suggests an interrelated hierarchy of that the vertical and horizontal complexity of limiting factors operating either individually, foliage do not correlate well with functional in an additive fashion, or synergistically to diversity of small mammals (MacMahon 1976, elicit a series of responses from the impacted Grenat and Serranos 1982). rodent population. Johnson and Hansen (1969) found that Previous studies of critical factors suggest 2,4-D (2,4-dichlorophenoxy acetic acid) her- that bicide treatments did not eliminate cover pro- population density is a consequence of survival, repro- vided by stems and branches of shrubs killed duction, and dispersal of individuals. These processes by the treatment. They observed no signifi- are affected by the capacity to tolerate expressions of the physical environment, the availability of essential cant differences in deer mouse density be- resources and interactions with other individuals of the tween treated and untreated . same and different species (Brown and Munger 1985). The opportunity to further explore the rela- They have shown that limited food resources tionships between small rodents and distur- as well as interspecific competition within a bances in their habitats came as a companion guild of ecologically similar species play a ma- study of habitat alterations designed to benefit jor role in regulating the density of desert the Utah prairie dog (Cynomys parvidens), a rodent populations. threatened species (U.S. Fish and Wildlife Other studies have indicated that habitat Service, 49CFR2330). This suite of habitat- manipulations adversely affect rodent food disturbing treatments, along with control

Department of Botany and Range Science, Brigham Young University, Provo, Utah 84602. Wildlife and Range Resources Laboratory, 407 WIDB, Brigham Young University, Provo, Utah S4602. ^Department of Zoology. 167 WIDB, Brigham Young University, Provo, Utah 84602. ^Department of Range Science, Texas A&M University, College Station, Texas 77843.

435 436 Great Basin Naturalist Vol. 49, No. 3

200

cd c o 00 -100 2 o Q. Q. O E CD H

Fig. 1. Climatic diagram representing weather variables (temperature, C; precipitation, mm) measured at Bryce Canyon, Utah, during 1986.

areas, allowed examination and testing of the tional cold desert shrubland. Dispersion of following hypotheses relating to small mam- vegetation varies from open grassy areas to mals (excluding C. parvidens): stands of closely or widely spaced shrubs.

1. Population densities of small rodents on treated plots Dominant shrubs in this area are black sage- would be reduced by a combination of factors includ- brush {Artemisia nova), fringed sagebrush ing depressed reproduction, increased dispersal, and (Artemisia frigida), carpet phlox (Phlox increased mortality. hoodii), dwarf rabbitbrush (Chrysothamnus 2. The size of home ranges of the various rodent species depressus), rubber rabbitbrush (Chryso- would generally increase on treated areas. thamnus nauseosus), winterfat (Ceratoides 3. indicator of body condi- Body weights of rodents, as an lanata), and broom snakeweed ( tion, would be proportionately less on treated plots. sarothrae). Dominant grasses are western 4. species would decrease in rela- Diversity of rodent wheatgrass (Agropyron smithii) and squirrel- tionship to vegetational treatments; interspecific tail (Elymus elymoides). Prominent forbs are dominance by one or more species may be evidenced. pretty rockcress (Arahis pulchra), vernal 5. The effects of vegetational treatments can be quantita- daisy (Erigeron pumilus), and longleaf phlox tively measured and expressed in ways meaningful to resident rodent populations. (Phlox longifolia). Scientific names of plants are after Welsh et al.(1987). Average monthly precipitation was 39 in 1986 but is highly Study Area and Methods mm variable (0 to 130 mm) from month to month. The study site (2.56 square km) is just west The wettest months of the year are from July of Widtsoe Junction, Garfield County, near to September. Mean monthly temperatures Bryce Canyon National Park. It was estab- range from —2 C in January to 17 C in July lished on lands administered by the state of (Fig. 1).

Utah in what is known as John's Valley. The Studies were conducted on 12 plots, average elevation at the study site is 2,290 m, each 165 x 165 m and located at least 0. 16 km and, although the terrain is relatively flat, the from adjacent plots. The experimental design northeast corner is dissected by a dry water- represented a randomized complete block course. Vegetation is primarily upper eleva- with three replications. Three treatments July 1989 Zou etal.: Rodent Habitat Manipulations 437

Fig. 2. The location of the study site in southern Utah and the spatial relationship of the treatment and control plots.

(September and October 1985) included (1) Each plot contained a 10 x 10 station trap- mechanical shredding (rotobeating) followed ping grid with each trap 10 m apart; thus, 100 by reseeding; (2) spraying with 3.3 kg/ha live-capture traps (38 x 11 x 11 cm) were 2,4-D (2,4-dichlorophenoxy acetic acid) per monitored in every plot each day of trapping. plot; and (3) spraying with 3.3 kg/ha 2,4-D per Trap stations were permanently marked with plot followed by reseeding (Fig. 2). The 12 a numbered stake. Traps were set on each plot study plots were divided into three blocks for five consecutive nights during three trap- with three treatment plots and one control ping periods (5-23 May, 23 June-11 July, plot in each block. All treated plots were en- 4-22 August 1986). Traps were baited with a closed with electrical wire fencing in May grain mixture after 6 p.m. and checked before 1986 to prevent grazing by livestock. 9 a.m. the next morning. This time interval 438 Great Basin Naturalist Vol. 49, No. 3 allowed capture of diurnal as well as nocturnal Population dynamics of the deer mouse, rodents. Trapped mammals were identified pocket mouse, and other rodents were inves- and checked for sex, age (adult or juvenile), tigated between May and August 1986. Home weight, reproductive condition, and station range size of the deer mouse was estimated number where trapped. They were then using the minimum area method (Dalke 1942, marked by toe-clipping before their release. Mohr 1947, Stickel 1954). Only those animals Trap nights were shifted slightly in August to captured three or more times with no more avoid the rodent mortality expected when than two captures on the edge of the grids trapping during wet weather. were included in calculations of home range. Data size of ranges quantita- Density and projected canopy cover of on home were tively transformed (Woolf 1968) to stabilize shrubs on the treatment and control plots the variance among samples from treatments were measured by the point-quarter method and control plots. A two-way ANOVA was (Mueller-Dombois and Ellenburg 1974). Veg- applied to these data as a test for differences in etation biomass in each plot was estimated home range size between deer mice among using a double sampling technique (Tadmor treatment and control plots during the trap- et al. 1975). Vegetation sampling occurred ping periods. within a series of 30 randomly selected 1-m quadrats (10 each on random lines) in each plot. There were three periods of vegetation Results sampling from May to September: (1) May- A total of 18,000 trap nights yielded 459 July, July-early August, (3) late August- (2) unique rodents for a trap success of 6.0%. September. Vegetation was sampled concur- Five species of rodents were caught; these rently with trapping. The difference between included the deer mouse (Peromyscus manic- biomass of shrubs, forbs, and grasses (and ulatus), 76.5%; Great Basin pocket mouse grasslike plants) among treatment and control (Perognathus parvus), 13.7%; sagebrush vole plots was tested using two-way (Ott ANOVA (Lagurus curtatus), 2.6%; Ord's kangaroo rat vegetative composition in 1984). Similarity of (Dipodomys ordii), 1.3%; and the diurnal each plot was analyzed using a method based least chipmunk (Eutamias minimus), 5.8%. on the frequencies of observed species Deer mice are widely distributed and were (Sneath and Sokal 1973). Regression analyses caught on all plots. Minimal data were gener- were used to investigate the relationship be- ated on the pocket mouse and other rodent tween diversity of vegetation and diversity of species; thus, statistical analyses, to associate rodents (MacArthur 1972). the effects of habitat manipulation on popula- Population density of desert rodents was tion characteristics, were confined to the deer estimated from mark recapture data using mouse. The other species noted above were Hayne's (1949) modified ratio method. This more restricted in distribution and were not method assumes no individual mortality, emi- present on all plots either because of overall gration, immigration, or trap avoidance by the low levels of population density or perhaps rodents under study during the trapping peri- as a result of specific treatments. For exam- ods. Seasonal changes are expected in popula- ple, least chipmunks were not captured on tion densities of desert rodents; therefore, the rotobeaten plots. Estimates of popula- separate estimates were generated in each tion density were derived for each trapping

trapping period. The averages of the three period during the seasons of sampling (Table 1,

estimates of population density in each treat- Fig. 3). ment and control plot were used as summary For the deer mouse population, nonpara- data. metric statistical analysis suggested no signifi- Those data derived from trapping rodents cant difference among treatments and control

that did not follow a normal distribution plots within each trapping season (Table 2). were analyzed using nonparametric statistical The deer mouse increased on plots from May methods to evaluate difference among treat- to August. Population increases were also ob- ment and control plots. The Kruskal Wallis served for other rodent species, but the num- test was used for paired comparisons within ber of pocket mice did not change significantly

each trapping period (Siegel 1956). among the three trapping periods (Fig. 4). In July 1989 Zou et al.: Rodent Habitat Manipulations 439

Table 1. Estimates of mean population density (number per hectare x ± SD) of five rodent species in relationship to habitat treatment and control plots in southern Utah. 440 Great Basin Naturalist Vol. 49, No. 3

| Deer mouse

Population size per hectare July 1989 Zou etal.: Rodent Habitat Manipulations 441

20 -

18 ..

16 .,

14 .. Density

12 ..

Pocket mouse 10 ..

8 .. O- Deer mouse

6 ., • 5 species

4 ..

2 ..

._ i

May 1986 June 1986 July 1986 August 1986

Collection Dates

Fig. 4. Seasonal trend in estimated population density plotted for the deer mouse and Great Basin pocket mouse; values for the other three rodent species are combined.

Table 3. Differences in mean size of home ranges of treatments. (2) Although vegetation fre- the deer mouse on plots that received various habitat quency varied among the treatment and con- treatments in Garfield County, Utah. trol plots, there was usually at least 50% treated plots Home range agreement. (3) Vegetation on the Treatment (x ± SD) was more homogeneous and had greater con-

2 gruity than vegetation control plots. Control 451 m ± 416*a on Rotobeating and reseeding 1110 m 2 ± 1144b Regression analysis suggests there was an 2 Herbicide 1177 m ± 863b inverse linear relationship (r = —.76) be- 2 Herbicide and reseeding 709 m ± 759a tween the diversity of the frequency of occur- *Values followed by different letters, a and b, are significantly different. rence for vegetation and the same diversity of

rodents on the study plots (Table 9). Grass biomass significantly increased on the = = herbicide treated plots (F 79.18, d.f. 2,4 Discussion p = .001), and this too was expected (Johnson 1964). There were no significant differences in Because biomass of graminoides and annual forb biomass between the treatment and con- forbs was less affected by habitat treatments trol plots (Table 5). than was biomass of shrubs, both grasses and Percent frequency of occurrence for com- forbs were more uniformly available as a food mon plant species on treatment and control source for rodents on treatment and control plots is shown in Tables 6, 7, and 8. Summa- plots. Biomass of grasses increased on plots rized data on frequency of occurrence for veg- treated by herbicide. Variation in plant spe- etation were submitted to cluster analysis cies diversity on plots fails to account for ro-

(Sneath and Sokal 1973) (Fig. 6), and the fol- dent species diversity because the most vege- lowing conclusions seem evident: (1) Seasonal tatively diverse plots had relatively lower changes in vegetation frequency were more rodent diversity (Table 9). This observa- evident than changes attributed to habitat tion questions the relationship between plant 442 Great Basin Naturalist Vol. 49, No. 3 July 1989 Zou etal.: Rodent Habitat Manipulations 443

Table 6. Percent frequency of occurrence for plant species during the first sampling period (May-early July) for treatment and control sites.

Species 444 Great Basin Naturalist Vol. 49, No. 3

Table 8. Percent frequency of occurrence for plant species during the third sampling period (late August- Septemher) for treatment and control sites.

Rotobeating Herbicide and Species Control and reseeding Herbicide reseeding

Grasses Agropyron smitJiii 70 93 Sitanion hystrix 30 Shrubs A rtemisia frigida 36 Artemisia nova 91 Artemisia pygmaea 8 Ceratoides lanata 60 Chrysothamnus depressus 16 Chrysothamnus nauseosits 7 Gutierrezia sarothrae 43 Phlox hoodii 92 Forbs Arabis pidchra Erigeron ptimilus 22 Phlox longifolia 3 1

July 1989 Zouetal.: Rodent Habitat Manipulations 445

Table 9. Values for an index of diversity (MaeArthur and this would be most important for very index, 1972) pertaining to frequency of occurrence for small mammals occupying high, cold desert vegetation and five rodent species on plots studied in relationship to habitat treatments. communities. Loss of functional cover would also increase the risk from avian and terres- Diversity of Diversity of trial nocturnal predators. Pocket mice are cre- Treatments vegetation rodents puscular and nocturnal and, thus, would be Control 9. 1.3 active concurrently with deer mice. Appar- Rotobeating and reseeding 6.5 1.6 ently the combination of factors associated Herbicide 6.6 1.5 with the prescribed habitat treatments nega-

tivelv impacted these pocket mice (Table 1, that reduce this third dimension of area avail- Figs' 3, 4). able for foraging may force resident deer mice A certain threshold of shrub canopy cover is to expand their two-dimensional home range. important for environmental amelioration for Rotobeating of shrubs would immediately re- the diurnal least chipmunk (Parmenter and duce the canopy aspects of home range, while MacMahon 1983) and for protection from treatment with 2,4-D would have a more avian and terrestrial predators. The severe gradual effect. The arthropod population de- reduction in shrub canopy cover on the roto- pendent on the shrub canopy would also be beaten plots probably induced least chip- eliminated, thus forcing deer mice to forage munks to emigrate to more favorable sites. for terrestrial arthropods, succulent vegeta- They remained on the herbicide-treated plots tion, and seeds. This behavior would bring until the third trapping period. Apparently them into more direct conflict with resident the dead canopy of shrubs provided adequate pocket mice over use of these essential re- cover until daily ambient temperatures sources. Data indicate deer mice apparently reached their highest levels, thus forcing dis- dominate in these kinds of interactions. persal (Table 1). The fossorial behavior of Brown and Munger (1985) showed that com- least chipmunks (Lechleitner 1969) would petition for limited food resources had a de- substitute for some loss of shrub canopy cover cided effect on organization of desert rodent but would not protect these rodents while communities. Species occupying different foraging. feeding and habitat guilds would not be ex- Sagebrush voles were uncommon and pected to respond in a competitive manner found only on rotobeaten plots during the first (Hallet et al. 1982). Deer mice readily forage sampling period after treatment, thereafter for seeds of grass, forbs, and shrubs (Everett being confined to the control plots. They have et al. 1978) and thus have seriously con- been observed to be active 24 hours of the tributed to failures of seedings day, year-round, and are considered strictly (Howard 1950, Spencer 1954, Nord 1965, vegetarian (Carroll and Genoways 1980), pre- Nelson et al. 1970). Their contributions to the ferring succulent vegetation such as tender failure of the seeding efforts in this study are leaves, nodes, young culms, and seeds in the unknown, but population levels on the herbi- dough stage (Maser et al. 1974). Sagebrush cide and reseeded plots remained near that of (usually Artemisia tridentata) leaves are the control plots. heavily utilized from October through Janu- The rather severe habitat treatments nega- ary; furthermore, sagebrush and rabbitbrush tively affected resident pocket mice (Table 1, (Chrysothamnus sp.), either alive or dead, Fig. 3). Following habitat treatments, interac- function as essential year-round cover (Maser tions between pocket mice and deer mice for et al. 1974). There is a negative relationship essential, but reduced, resources may have between wind velocity at ground level and registered negative impacts on pocket mice. activity of sagebrush voles aboveground

The mechanical disturbances of rotobeating (Maser et al. 1974). Therefore, treatments and, to a lesser extent, herbicide treatment that reduced the aboveground vegetation may have collapsed the shallow network of would contribute to increased wind at the tunnels usually maintained by the Great Basin ground surface. The prescribed habitat treat- pocket mouse (Banfield 1974). A dense crown ments were detrimental to the short-term and cover of vegetation prevents chilled air from long-term requirements of sagebrush voles settling to ground level at night (Smith 1966), (MaeArthur 1972). 446 Great Basin Naturalist Vol. 49, No. 3

The overall plant community and rather tion density of the deer mouse and Great heavy clay to stony soils on the study site are Basin pocket mouse did not support this as- probably not ideal habitat for Ord's kangaroo sumption. Scanty data regarding least chip- rats. These nocturnal rodents are rather poor munks and sagebrush voles seemed to agree burrowers because of their weak forelegs and with this premise. Seasonal trends indicated slender claws and are usually found in burrow that reproduction of pocket mice and perhaps systems dug in loose sand or loess (Banfield survival of juveniles were negatively affected 1974). The process of rotobeating on plots by the prescribed habitat perturbations. often uprooted shrubs from the soil; this ac- Hypothesis 2: The size of home ranges of tion created soil conditions more compatible the various rodent species would generally with the requirements of kangaroo rats. The increase on treated areas. Trapping results more xeric conditions created by this treat- provided significant data only for the deer ment were probably not detrimental to this mouse, and home ranges of these mice in- desert-adapted species. creased in size apparently in response to habi- This study did not address other factors that tat treatments. could and, perhaps, did influence the popula- Hypothesis 3: Body weights of rodents, as tions of small mammals on the treatment and an indicator of body condition, would be pro- control plots. Beneficial data and insight could portionately less on treated plots. Adequate be gained from documenting direct behav- data on the deer mouse did not validate this ioral interactions between species, determin- hypothesis. Habitat treatments did not limit ing the effects of disease and parasites, ob- food or the foraging capabilities of this oppor- serving differential responses of predators, tunistic mouse. looking at unresolved mortality, documenting Hypothesis 4: Diversity of the rodent spe- long-term reproduction and dispersal, and cies would decrease in relationship to vegeta- examining ecological relationships with the tional treatments; interspecific dominance by Utah prairie dog. one or more species may be evidenced. Lack Funding constraints did not allow for a of significant data on rodents other than the longer period of posttreatment investigation. deer mouse weakens the conclusion that this We concur with Wiens and Rotenberry (1985) premise is not true (r = —.87). The deer in appreciating that time lags in responses of mouse appears to be interspecifically domi- individual animals on treated plots, as well as nant in its guild and probably most affected behavioral fidelity to sites regardless of treat- the Great Basin pocket mouse on treated ments, make data from these kinds of studies sites. more difficult to interpret. Hypothesis 5. The effects of vegetational treatments can be quantitatively measured Conclusions and expressed in ways meaningful to those who investigate resident rodent populations. Evaluation of field data in relationship to Measured reductions in canopy cover of the working hypotheses provides useful in- shrubs on treated plots agreed with two- sight for ecological studies that use experi- dimensional expansion of home ranges of the mental designs involving evaluations of treat- deer mouse. The canopy of shrubs serves as a ment and control data. Results in our study third-dimension of foraging habitat for the were compromised somewhat by restricting deer mouse, and they respond to removal of the evaluation to the first season after habitat this resource. treatments, failure of the reseeding effort, and confounding by factors similar to those dis- Acknowledgments cussed by Wiens and Rotenberry 1985. A concise evaluation of these premises would We thank personnel of the Non-Game include the following. Wildlife and Game Management sections of Hypothesis 1: Population densities of small the Utah Division of Wildlife Resources for mammals on treated plots would be reduced significant monetary support and experimen- by a combination of factors including de- tal counsel. Appreciation is extended to the pressed reproduction, increased dispersal, U.S. Bureau of Land Management for use of and increased mortality. Our data on popula- comfortable living facilities while in the field. July 1989 Zou etal.: Rodent Habitat Manipulations 447

Brigham Young University also provided criti- Crooked River National , Jefferson County, Oregon: a contribution to its life history cal monetary support. Dr. Samuel R. Rush- and ecology. Saugetierk. Mitt. 22: 193-222. forth provided expert counsel regarding clus- Mohr, C O 1947. Table of equivalent populations of ter analysis. Gratitude is extended to James North American small mammals. Amer. Midi. Mackley for helping with many aspects of the Nat. 37: 223-249.

eller-Dombois, ., research and to Dale Thurber for assistance in Ml D AND H. Ellenburg 1974. Aims and methods of vegetation ecology. Wiley, N.Y. trapping small mammals. J. 545 pp.

Nelson J R A M Wilson, and C J Goebel 1970. Literature Cited Factors influencing broadcast seeding in bunch-

grass range. J. Range Manage. 23: 163-170. Banfield, A W. F. 1974. The mammals of Canada. Uni- Nord, E C 1965. Autecology of bitterbrush in California. versity of Toronto Press, Toronto. 438 pp. Ecol. Monogr. 35: 307-334.

Brown, J H . and J C MUNGER. 1985. Experimental O'Neal, G T . J T Flinders, and W P Clary 1987. manipulation of a desert rodent community: food Behavioral ecology of the Nevada kit fox (Vulpes addition and species removal. Ecology 66: macrotis nevadensis) on a managed desert range- 1545-1563. land. Pages 443-481 in H. H. Genoways, ed.,

Carroll, L. E., and H. H Genoways. 1980. Lagurus Current mammalogy. Vol. 1. Plenum Publ. curtatus. Mammal. Species 124: 1-6. Corp.. N.Y. 519pp. Dalke, P D. 1942. The cottontail rabbits in Connecticut. Ott. L. 1984. An introduction to statistical methods and State Geol. and Nat. Hist. Surv., Bull. 65: 1-97. data analysis. Pages 636-655 in PWS Publ.,

Everett, R L , R O Meeuwig, and R Stevens 1978. Boston, Massachusetts.

Deer mouse preference for seed of commonly PARMENTEB B R. \\i) J A MacMahon 1983. Factors planted species, indigenous weed seed, and sacri- determining the abundance and distribution of

fice foods. J. Range Manage. 31: 70—74. rodents in a shrub steppe ecosvstem: the role of FLAKE, L. D. 1973. Food habits of four species of rodents shrubs. Oecologia 59(2-3): 145-156.

on a short-grass prairie in Colorado. J. Mammal. Rosenzweig, M L. B Smigel, \\d A Kraft 1975. Pat- 54: 636-647. terns of food, space and diversity. Pages 241-268 Grenat, C, and V. Serrano 1982. Spatial partitioning in P. K. Ghosh, ed., Bodents in desert environ- and structure of small vertebrate communities in ments. Dr. W. Junk, The Hague.

the Chihuahuan Desert. C. R. Soc. Biogeogr. 58: Rosenzweig M L., and J. Winakub 1969. Population 159-191. ecology of desert rodent communities: habitats

Hallett, J G , M A. O'Connell, and R L Honeycutt. ami environmental complexity. Ecology 50: 1982. Competition and habitat selection test of a 558-572. theory using small mammals. Oikos 40: 175-181. Siegel. S 1956. Nonparametric statistics for the behav- Hayne, D W 1949. Two methods for estimating popula- ioral sciences. McGraw-Hill, N.Y. 312 pp.

tion from trapping records. J. Mammal. 30(4): Sneath, P H A , and R R SOKAL, 1973. Numerical tax- 399-411. onomy. W. H. Freeman and Co., San Francisco.

Holbrook, S J 1979. Species diversity patterns in some 573 pp. present and prehistoric rodent communities. Smith, R L 1966. Ecology and field biology. Harper and Oecologia 44: 355-367. Row, N.Y. 686 pp.' Howard. W. E. 1950. Wildlife predations in broadcast Spencer. D. A 1954. Rodents and direct seeding.

seeding of burned brushlands. J. Range Manage. J. Forest. 52: 824-826. 3: 291-298. Stickel, L. F. 1954. A comparison of certain methods of

Johnson, D. R 1964. Effects of range treatment with measuring ranges of small mammals. J. Mammal. 2,4-D on food habits of rodents. Ecology 45: 35: 1-15.

341-349. Tadmor, N H ., A Griecher. I Noy-meir, R W Ben-

Johnson, D. R , andR M Hansen 1969. Effects of range jamin, and E. Eyal. 1975. An evaluation of the treatment with 2,4-D on rodent populations. calibrated weight-estimate methods for measuring

J. Wild. Manage. 33: 125-132. production in annual vegetation. J. Bange Man- KoTLER, B. P 1984. Risk of predation and the structure of age. 28: 65-69.

desert rodent communities. Ecology 65: 689-701. Taitt, M. J. 1981. The effect of extra food on small rodent Lechleitner, R. R 1969. Wild mammals of Colorado. populations. I. Deermice (Peromyscus manicula-

Pruett Publishing Co. , Boulder, Colorado. 254 pp. tus). J. Anim. Ecol. 5: 111-124.

MacArthur, R H. 1972. Geographical ecology: patterns Taylor, R J 1963. Vertical distribution of Peromyscus in the distribution of species. Harper and Row leucopus and P. gossypinus under experimental Press, N.Y. 269 pp. conditions. Southwest Nat. 8: 107-115.

MacMahon. J A 1976. Species and guild similarity of Thomson, R. D 1981. Spring and summer diets of desert North American desert mammal faunas: a func- rodents: food resource partitioning in the Uinta tional analysis of communities. Pages 134-148 in Basin. Unpublished thesis, Brigham Young Uni- D. W. Goodall, ed.. Evolution of desert biota. versity, Provo, Utah. University of Texas Press, Austin. Wecker, S. C 1963. The role of early experience in habi- Maser, C. E., W. Hammer, C Brown, R E Lewis, tat selection by the prairie deer mouse, Per- R L Rausch, and M. L. Johnson. 1974. The omyscus maniculatus Bairdi. Ecol. Monogr. 33: sage vole, Lagurus curtatus (Cope, 1868), in the 307-325. .

448 Great Basin Naturalist Vol. 49, No. 3

Wolff, O 1984. The effects of density, food, and inter- Welsh S. L.. N D Atwood. L C Higgins, and J specific interference on home range size in Pero- S. Goodrich. 1987. A Utah flora. Great Basin Nat. leucopus and Peromyscus maniculatus Mem 9.894 pp. myscus 2657-2662. 1985. Response of Canadian J. Zool. 63: WlENS, A., AND J. T. Rotenberry. J. biometry. D. Van Nos- breeding passerine birds to rangeland alteration in Woolf, G M. 1968. Principles of

trand Co. , Inc. , Princeton, N.J. 359 pp. a North American shrubsteppe locality. J. Applied Ecol. 22: 655-668.