Abstract.--Investigations of vegetation in Wichita County, Texas
indicate that changes in patterns of grazing and the introduction of
non-native plant species may affect populations of the Texas kangaroo
rat. Intensely and moderately grazed areas were compared to each other
and to a previous investigation involving an ungrazed pasture dominated
by introduced Japanese brome (Bromus japonicus). Thirty Dipodomys elator
were trapped at the intensely and moderately grazed sites, whereas only
two animals were caught on the periphery of the ungrazed site in Wichita
County. In addition, the moderately grazed site was compared to the
intensely grazed site and no significant differences in vegetative
richness or percentages grass and forforb were found between sites.
Height of vegetation, percentage bare ground and woody species coverage
were significantly different in comparisons between the two grazed
sites. Because the two sites contained populations of D. elator, it
appears that they can use moderately to heavily grazed habitats as
burrow locations and can tolerate significant differences in vegetation
height and amount of bare ground and woody vegetations. They rarely use
ungrazed sites as habitat and, in a previous investigation, an ungrazed
site was significantly different from the grazed sites in vegetational
height, percentage bare ground, and percentage grass coverage. Grazing
regimes, amount of bare ground coverage, and introduction of tall,
dense-growing grasses may be important considerations in managing
habitat for Texas kangaroo rats. Moderately to heavily grazed sites may
provide better habitat for these state-threatened mammals.
Dipodomys elator (Merriam 1894) is a state threatened species. The
International Union for the Conservation of Nature (1986) lists habitat
loss and degradation resulting from agricultureagricultural and
infrastructure development as major threats. Though degradation such as
fragmentation and loss of habitat have played important roles, changes
in vegetation patterns may also be important. Much historic range of the
Texas kangaroo rat has been fragmented by extensive cultivation within
the Rolling Plains Region of Texas and adjacent regions of Oklahoma
(Correll & Johnston 1970).
Cultivation fragmented the grasslands and only those areas
unsuitable for cultivation were left in their natural state. These
fragments were fenced by ranchers and grazed by cattle and provided some
habitat for the Texas kangaroo rat. However, because of a decline in
ranching, some of these pasturelands are no longer grazed and have been
invaded by introduced species such as Japanese brome. Additionally, lack
of fire has allowed the increase of woody vegetation such as honey
mesquite (Prosopis glandulosa) and lotebush (Ziziphus obtusifolia). In
some cases, Texas kangaroo rats used these woody species as burrow sites
because the plants collected wind-blown soil and the Texas kangaroo rats
dig burrows at their bases. However, as mesquites mature, their shade
changes the vegetation composition sometimes favoring introduced grasses
like Japanese brome, which grows densely and changes the habitat so that
it is more suitable for other types of small mammals.
Mesquite forest was not seen as a problem by the Texas Parks and
Wildlife Department (TPWD) because of the assumption that this habitat
was required for Texas kangaroo rats. However, perceptions regarding
threats to the species and ideas about future management have changed.
Research has suggested that (1) mesquite is not a critical component of
D. elator habitat (Stangl et al. 1992; Goetze et al. 2007), (2) grazing
may benefit D. elator (Stangl et al. 1992; Stasey 2005; Goetze et al.
2007), and (3) Texas kangaroo rats opportunistically use human
structures that collect friable soils as burrow sites (Stangl et al.
1992; Stasey 2005; Goetze et al. 2007). Schmidly (2004), which is used
by TPWD as their main source of small mammal data for the state, states
that heavily grazed rangeland and the eroded sites of rangeland roadways
may provide optimum habitat. NatureServe (2008) states that vegetation
has become overgrown and that rangeland practices that result in dense
growth of grasses or invasion of non-native grasses have degraded
habitat because the Texas kangaroo rat thrives in heavily grazed or
otherwise disturbed conditions. NatureServe (2008) further states that
habitat for D. elator consists of sparsely vegetated areas that may or
may not include honey mesquite, including heavily grazed land, disturbed
areas, and areas along fencerows adjacent to cultivated fields and
roads. These current statements are quite different from Davis &
Schmidly (1994), who stated that Texas kangaroo rat burrows invariably
entered the ground at the base of a mesquite and the primary threat
contributing to the rarity of the species was the clearing of mesquite
Currently, the only relatively large populations of Texas kangaroo
rats known in Wichita County occur in pastures with small or scattered
mesquite, and burrows are often not associated with mesquite at all, but
rather with lotebush, prairie mounds (natural, elevated, and relatively
bare areas possibly uplifted by clay soils swelling in cracks; Diggs et
al. 1999), or in areas where man-made berms occur due to road, fence,
and oilfield construction, or in association with old (>30 years),
unburned brush piles where wood has decayed leaving a mound of loose
friable soil (Stangl et al. 1992; Goetze et al. 2007). Stangl et al.
(1992) hypothesized that grazing bison and prairie dogs, along with
fire, historically maintained the type of disturbances needed by the
Texas kangaroo rat. Also, prolonged drought likely played an important
role in fire frequency and maintaining short vegetation with
intermittent bare patches of soil. Natural prairie heterogeneity such as
prairie mounds (Diggs et al. 1999; Goetze et al. 2007) appear to be
important in providing the type of habitat needed by the Texas kangaroo
rat before cattle grazing and human mediated disturbances were used
opportunistically as burrow sites (Stangl et al. 1992). In areas where
cattle no longer graze or at sites where native vegetation has been
replaced by introduced species, it appears that populations of Texas
kangaroo rats have declined.
Ecological characterization of burrows in situations that lacked
grazing as a component are rare (Martin & Matocha 1992; Stasey
2005). Martin & Matocha (1992) trapped for 10 trap nights and
characterized a burrow where a single Texas kangaroo rat was trapped in
association with a fence row adjacent to a gravel county road in Motley
County, Texas. No other mammals were captured. This capture site
contained vegetation characteristic of a disturbed site and was adjacent
to a field of sudan grass. At the capture site there was 30.2% bare
ground, 65% grasses and 4.8% forb (Martin & Matocha 1992).
Stasey (2005) trapped for 972 trap nights in a mesquite forest with
an understory dominated by Japanese brome. His site was within 4.0 km of
a large and persistent Texas kangaroo rat population in Wichita County
where the grazing regime is heavy (Goetze et al. 2007). During his
investigation, he caught only two kangaroo rats on the periphery of the
mesquite forest habitat. One was caught in a friable clay soil that had
blown in and accumulated around the corner posts that supported the gate
leading into the pasture and the other was captured along the fence
separating this pasture from a wheat field where a berm had accumulated
due to plowing next to the fence line. Stasey (2005) caught Sigmodon
hispidus, Peromyscus leucopus, and P. maniculatus in the core of the
ungrazed site. In the core of the ungrazed pastureland, where no Texas
kangaroo rats were captured, seven quadrats were sampled to assess
vegetation characteristics. Percent bare ground had a mean of 10.9%,
grasses 63.7%, and forbs 16.1%, whereas the mean average herbaceous
height was 49.0 cm (Stasey et al. 2005).
Because identifying habitat critical to the survival of the species
is a research priority (Jones et al. 1988), the purpose of this
investigation was to compare a moderately grazed site to that of a
heavily grazed site, both of which have populations of kangaroo rats.
These data were then compared to data from ungrazed sites dominated by
Japanese brome (Stasey 2005) and vegetation associated with a disturbed
roadside (Martin & Matocha 1992).
MATERIALS AND METHODS
The study area is in Wichita County on the east side of the
intersection of highways 1739 and 2384 and is a privately owned ranch
that is moderately grazed pasture (0.30 head per ha). Coordinates at the
entrance of the ranch are 34.05423 N, 98.81721 W. The pastureland is
fenced, has small mesquite (less than 2 m. in height), and has several
old oil field storage sites. Mesquite density at the site is 168/ha.
This site was compared to a nearby population of Texas kangaroo rat
known in Wichita County and its locality and history have previously
been described (Stangl et al. 1992; Goetze et al. 2007). Grazing at this
site is intense (0.81 head/ha) and it has about 54 small (less than 2m
in height with most under 1 m) mesquite/ha (Goetze et al. 2007).
All trapping was done using 7.5 by 8.8 by 30 cm Sherman traps with
rolled oats as bait. Traps were set just before dark and checked early
the next day. Based on parameters set by Stangl et al. (1992) and Stasey
(2005) regarding burrow entrance diameter, angle of entry, and
vegetation, suspected burrows were selected at the sites and three traps
were placed around each burrow entrance.
All vegetation data was quantified in May so that direct
comparisons could be made, thus eliminating seasonal vegetative changes.
For burrows where at least one Texas kangaroo rat was caught, one square
meter quadrats were centered around burrow entrances and percentage
cover, grass, forb, bare ground, and woody vegetation (when present) was
recorded (Goetze et al. 2007). Vegetative richness and height were
measured and the dominant grass, forb, and woody species of each quadrat
were identified (Goetze et al. 2007). Quantitative data was compared
using SPSS 14.0 (SPSS, Inc. 2005). A Mann Whitney test was used to test
for significant differences in richness and percentages of grasses,
forbs, bare ground, and woody vegetation between the two grazed sites.
In addition, data from a previous study (Stasey 2005) at a site that was
ungrazed was included in a Kruskal-Wallis analysis to test for
significant differences between the grazed and ungrazed sites.
Dominant vegetation at the site was identified using floras for the
state and for north central Texas (Correll & Johnson 1970; Diggs et
al. 1999). Voucher specimens are deposited in the Tarleton State
University Herbarium (TAC).
The specific location of each burrow was recorded in decimal
degrees using a Garmin GPS-12 unit and burrows were classified as being
associated with human-mediated disturbances such as old brush piles and
fence rows, or other available natural habitats that included prairie
mounds, which are elevated, open areas formed by clay soil shrinkage and
swelling (Diggs et al. 1999), or accumulation of soil at the base of
lotebushes or honey mesquites (Goetze et al. 2007; Table 1).
At the moderately grazed site, three traps were placed around each
of 12 burrows which resulted in the capture of 10 D. elator. No other
rodents were captured. Three of the animals were caught in each of the
traps placed at one burrow and two were caught in the three traps
surrounding another burrow site. The dominant grass at five of the seven
burrows was little barley (Hordeum pusillum) with buffalograss (Buchloe
dactyloides) and rescue grass (Bromus catharticus) being dominant at the
other two burrows (Table 1). The dominant forb at six of the seven
burrows was Texas broom weed (Gutierrezia texana) with Virginia
pepperweed (Lepidium virginicum) being dominat at the other burrow
(Table 1). Five of the seven quadrats had honey mesquite (Prosopis
glandulosa) as the dominant woody vegetation and two contained no woody
vegetation (Table 1). Two burrows were associated with fence rows and
five with honey mesquite (Table 1).
At the heavily grazed site, three traps were placed around each of
22 burrows which resulted in the capture of 18 D. elator. Of these, 10
burrow sites were analyzed for vegetation and burrow associations. At
the heavily grazed site, little barley was always the dominant grass and
most quadrats contained Virginia pepperweed as the dominant forb (Table
1). Other herbaceous dominants included Texas broomweed, hog potato
(Hoffmannseggia glauca), and western ragweed (Ambrosia psilostachya)
(Table 1). Woody vegetation was evenly distributed between lotebush and
honey mesquite. At the heavily grazed site, five burrows were associated
with old brush piles, two with prairie mounds, one with a fence row, and
one each with lotebush and honey mesquite (Table 1).
Average herbaceous height, percentage bare ground, and percentage
woody vegetation were significantly different between heavily and
moderately grazed sites (Table 2). Percentage forbs, grasses, and other
categories such as rocks and stumps, as well as richness were not
significantly different between sites (Table 2).
Average herbaceous height (P = 0.001), percentage bare ground (P =
0.009) and grasses (P = 0.009) were significantly different when
comparing the moderately and heavily grazed sites as well as the
ungrazed site examined by Stasey (2005).
Little barley, Texas broomweed, hog potato, Virginia pepperweed,
and western ragweed occur in disturbed habitats (Diggs et al. 1999).
These plants were dominant species associated with burrows of Texas
kangaroo rats and their occurrence is likely caused by disturbances such
as grazing of cattle and rodent activity around the burrows. Habitat of
D. elator was dominated by short, herbaceous vegetation (2.0 - 40.0 cm
in height). There is general agreement that D. elator requires a sparse,
short-grassland habitat (Dalquest & Collier 1964; Roberts &
Packard 1973; Carter et al. 1985; Stangl et al. 1992), and findings from
this current study support this conclusion. These findings also indicate
that grazing may be important in maintaining sparse, short grassland
habitat. When comparing two grazed sites and an ungrazed site (Stasey
2005), the only significant differences were in average herbaceous
vegetation height and percentage bare ground and grass. Grazing can
change these three parameters, which appear to be important in
maintaining Texas kangaroo rat habitat. As previously discussed, this is
complicated by the dominance of the introduced grass, Bromus japonicas
at the ungrazed site and additional studies need to be conducted at
ungrazed sites containing native vegetation. However, based on the lack
of D. elator at this site and the relative abundance at the grazed
sites, it appears that grazing plays a role in maintaining suitable
habitat for Texas kangaroo rats. Lack of grazing significantly increased
vegetation height at the ungrazed site, which on average was double the
moderately grazed site and seven times greater than the heavily grazed
site. Percentage of grasses was about three times greater on average at
the ungrazed site when compared to grazed sites. Percentage bare ground,
on average was about one-half that of the moderately grazed site and
about one-fourth that of the heavily grazed site. This tall, dense
coverage by grasses may impede Texas kangaroo rat movement, inhibit
their ability to see potential predators, and may make burrow
construction difficult. Lack of bare ground likely inhibits their dust
Burrows at the grazed site compare favorably to the burrow
ecological characteristics reported from Motley County (Martin &
Matocha 1992). In the classification system used for burrows and
described in the Materials and Methods, this burrow would have been a
fence line association and the animals were likely using soil that
accumulated at the base of the fence. Although no grazing was reported
at this site, the value reported for bare ground percentage (30.2%)
compares favorably with the grazed sites in this investigation but forb
percentage (4.8%) was low and grass percentage was high (65%). The
location of this burrow in a fence row berm may have provided the
friable soil preferred for burrow development. Its location at the edge
of a sudan field adjacent to a gravel road (Martin & Matocha 1992)
may have provided enough disturbance to maintain bare patches for dust
bathing and for foraging trails that still allow the rats to spot
Of the 17 burrows examined in this investigation, 47% were
associated with human disturbances including old brush piles and fence
rows. The other burrows were associated with more natural sites such as
shrubs and prairie mounds. In the heavily grazed site most burrows were
associated with 30 year old brush piles, while at the moderately grazed
site most were associated with honey mesquite. Second in number of
burrow associations at the heavily grazed site was prairie mounds and
none of this association was observed at the moderately grazed site.
Heavy grazing might make prairie mounds more suitable as burrow sites,
since that was the only place this habitat association occurred. At the
heavily grazed site, lotebush and honey mesquite associations were equal
in number while no lotebush associations were observed at the moderately
grazed site. This is likely because the moderately grazed site had woody
vegetation dominated by honey mesquite whereas the heavily grazed site
had some lotebush available. The type of shrub is probably not as
important as is the accumulation of loose, friable soil at the base of
Extrinsic disturbance caused by grazing, fire, or drought, and
natural landscape heterogeneity such as prairie mounds probably is
important for burrow site selection in Texas kangaroo rats. The slight
elevation of prairie mounds may provide more bare ground because of a
drier microclimate and better drainage. Also, the animals can dig their
more characteristically horizontal openings into the sides of these
mounds. Opportunistic use of any natural or manmade disturbance where
friable soil accumulates such as around shrubs, the bases of rocks,
fence lines, cattle pens, pasture and oil field roads, abandoned
equipment, old, unburned brush piles with most of the wood decayed have
been observed (Stangl et al. 1992; Goetze et al. 2007). Almost one-half
of the burrows in this investigation may be characterized as such,
supporting hypotheses of opportunistic use made by others (Stangl et al.
1992; Stasey 2005; Goetze et al. 2007). These manmade disturbances
likely mimic natural prairie heterogeneity.
Investigations of Texas kangaroo rats that examine symbiosis with
prairie dogs, effects of fire in maintaining habitat, the role of
drought on habitat, and additional research into the influence of
natural prairie heterogeneity and grazing regimes are critical for
understanding the animal's niche. Surveys need to be conducted at
Buffalo Creek Reservoir and Lake Arrowhead State Park to ascertain if D.
elator occurs in any protected natural areas (pnas) in Wichita County or
if not, if suitable habitat is available in protected natural areas.
Also, based on the results of this investigation, managers of pnas may
need to consider grazing as a management practice to promote the
development of habitat for the Texas kangaroo rat and other organisms
that require grazing as a disturbance.
We would like to thank Oscar and Edith Goetze for allowing access
to their properties in Wichita County and for room and board while
conducting fieldwork. Tarleton State University Organized Faculty
Research provided funding for parts of this project. This study was
conducted under Texas Parks and Wildlife permit SPR-0496-775.
Carter, D. C., W. D. Webster, J. K. Jones, Jr., C. Jones & R.
D. Suttkus. 1985. Dipodomys elator. Mammalian Species, 232:1-3.
Correll, D. S. & M. C. Johnston. 1970. Manual of the vascular
plants of Texas. Texas Research Foundation. Renner, Texas. 1083 pp.
Dalquest, W. W. & G. Collier. 1964. Notes on Dipodomys elator,
a rare kangaroo rat. Southwestern Nat., 9:146-150.
Davis, W. B. & D. J. Schmidly. 1994. The mammals of Texas.
Texas Parks and Wildlife Press. Austin, TX, 338 pp.
Diggs, G. M., B. L. Lipscomb & R. J. O'Kennon. 1999.
Shinners & Mahler's Illustrated Flora of North Central Texas.
Botanical Research Institute of Texas. Fort Worth, Texas, 1626 pp.
Goetze J. R, W. C. Stasey, A. D. Nelson, & P. D. Sudman, 2007.
Habitat attributes and population size of Texas kangaroo rats on an
intensely grazed pasture in Wichita County, Texas. The Texas J. of Sci.,
International Union for Conservation of Nature and Natural
Resources. 1986. 1986 IUCN red List of threatened animals. IUCN.
Cambridge, U. K., 105 pp.
Jones, C., M. A. Bogan & L. M. Mount. 1988. Status of the Texas
kangaroo rat (Dipodomys elator). The Texas J. of Sci., 40(3):249-258.
Martin, R. E. & K. G. Matocha. 1991. The Texas kangaroo rat,
Dipodomys elator, from Motley County, Texas, with notes on habitat
attributes. Southwestern Nat., 36:354-356.
Merriam, C. H. 1894. Preliminary descriptions of eleven new
kangaroo rats of the genera Dipodomys and Perodipus. Proc. Biol. Soc.
NatureServe. 2008. NatureServe Explorer: An online encyclopedia of
life [web application]. Version 7.0. NatureServe, Arlington, Virginia.
Available http://www.natureserve.org/explorer. (Accessed: September 4,
Roberts, J. D. & R. L. Packard. 1973. Comments on movements,
home range and ecology of the Texas kangaroo rat, Dipodomys elator
Merriam. J. Mamm., 54:957-962.
Schmidly, D. J. 2004. The mammals of Texas, revised edition.
University of Texas Press, Austin, Texas, 501 pp.
Stangl, F. B., Jr., T. S. Schafer, J. R. Goetze & W. Pinchak.
1992. Opportunistic use of modified and disturbed habitat by the Texas
kangaroo rat (Dipodomys elator). The Texas J. of Sci., 44(l):25-35.
Stasey, W. C. 2005. An evaluation of Texas kangaroo at (Dipodomys
elator): Biological habits and population estimation. Unpublished
Masters Thesis. Tarleton State University, 45 pp.
SPSS, Inc. SPSS 14.0 brief guide. 2005. Prentice Hall. Upper Saddle
River, New Jersey, 245 pp.
ADN at: firstname.lastname@example.org
Allan D. Nelson, Jim R. Goetze *, Elizabeth Watson and Mark Nelson
Department of Biological Sciences, Box T-0100, Tarleton State
University, Stephenville, Texas 76401 and * Science Department, Laredo
Community College Laredo, Texas 78040
Table 1. Dominant vegetation and burrow classifications at the two
study sites. Dominant forbs are broomweed (Gutierrezia texana), hog
potato (Hoffmannsegia glauca). pepperweed (Lepidium virginicum), and
ragweed (Ambrosia psilostachya). Dominant grasses are barley (Hordeum
pusillum), buffalo grass (Buchloe dactyloides), and rescue grass
(Bromus catharticus). Woody vegetation includes lotebush (Zizyphus
obtusifolia) and mesquite (Prosopis glandulosa). Burrow classifications
included in this table are defined in the methods section of this
Moderately grazed burrows:
Burrow # 1 2 3 4 5 6 7
Broomweed X X X X X X
Barley X X X X X
Mesquite X X X X X
None X X
Fence X X
Mesquite X X X X X
Heavily grazed burrows:
Burrow # 1 2 3 4 5 6 7 8 9 10
Hog potato X
Pepperweed X X X X X X
Barley X X X X X X X X X X
None X X X X X X X X
Old brush pile X X X X X
Prairie mound X X
Table 2. Comparison of heavily and moderately grazed sites for average
herbaceous height, percentage coverage of bare ground, forbs, grasses,
woody, other (rocks, stumps, posts, etc.), and richness. Comparisons
are made using means, standard deviations (in parentheses), and ranges
[in brackets] and evaluated using Wilcoxon Mann-Whitney test.
Significant differences at P < 0.05 are denoted by an asterick.
Heavily Grazed Moderately Grazed P-value
Avg. Herb. 7.1 ([+ or -] 7.9) 24.5 ([+ or -] 12.7) 0.0001 *
Height [2.0-40.0] [9.0-29.4]
% Bare Ground 49.9 ([+ or -] 24.0) 20.7 ([+ or -] 18.1) 0.024 *
% Forbs 16.5 ([+ or -] 13.1) 33.7 ([+ or -] 22.6) 0.133
% Grasses 24.60 ([+ or -] 18.9) 20.1 ([+ or -] 15.6) 0.623
% Woody 6.0 ([+ or -] 15.8) 26.1 ([+ or -] 26.5) 0.037 *
% Other 2.0 ([+ or -] 4.2) 0.0 ([+ or -] 0.0) 0.222
Richness 6.2 ([+ or -] 2.4) 7.6 ([+ or -] 1.8) 0.137