EDWARDS ET AL.— MAMMAL ABUNDANCE AT FORT HOOD 
3 
armadillos (. Dasypus novemcinctus ), lagomorphs, and 
large rodent species such as squirrels (Sciurus niger). 
The objectives of this research were to: 1) deter¬ 
mine medium-sized mammal species composition and 
calculate relative abundances of species present in high 
and low use areas within the three major selected habitat 
types, 2) determine any significant difference in relative 
abundances of medium-sized mammal species between 
areas of high and low military use within the three se¬ 
lected habitat types, as well as any differences existing 
due to seasonal, day of sampling, or transect effect, and 
3) determine the effectiveness of field survey methods, 
including live-trapping and nocturnal census counts, for 
monitoring the medium-sized mammal populations 
present. 
METHODS 
Fieldwork was conducted from September 1995 
through January 1997. During the first four months, a 
preliminary survey was conducted to select appropriate 
study sites. Three of Fort Mood’s four major plant com¬ 
munities were selected to serve as the major habitat types 
to be sampled during the survey. These include: riparian 
habitats consisting primarily of pecan, American elm, 
sycamore ( Platanus occidentalis), live oak ( Quercus 
fusiform is), and bigtooth maple (Acer grandidentatum ); 
upland habitats consisting of Ashe juniper, Texas oak 
(Quercus texana ), live oak, and netleaf hackberry (Celtis 
reticulata ); and savannah habitats consisting of little 
bluestem, yellow Indiangrass, sideoats grama, 
buffalograss (Buchloe dactyloides ), and common 
broomweed ( Xanthocephalum dracuncubides), with 
scattered Ashe juniper, and honey mesquite (Cornell and 
Johnston, 1979). 
Survey sites were selected on the basis of high and 
low military use within each habitat type following con¬ 
sultation with the Fort Hood Environmental Division and 
Fort Hood wildlife officials (Fig. 1). Designation of high 
and low use reflected the amount of military activity 
occurring regularly in the area. High military use areas 
are utilized regularly by troops, armored and/or tracked 
support vehicles. Low use sites are utilized primarily 
for the training of ground troops with little armored and/ 
or tracked vehicle disturbance. The six sites selected 
were surveyed to test whether high and low use areas 
were quantifiably different for degree of vehicle distur¬ 
bance. A 16 km transect was driven in each of the six 
survey areas. At 1.0 km intervals, a random compass 
direction was selected in which a 50 m transect was 
walked. The observer attached one end of a 100 m mea¬ 
suring tape at the beginning of the transect and walked 
50 m in the selected direction. Number of obvious roads 
or vehicle trackways crossed within the 50 m transect 
were recorded. A mean (total roads crossed) was calcu¬ 
lated for each area by taking the total number of roads 
crossed within each area and dividing by the total num¬ 
ber of transects sampled in each area. A student’s t-test 
was used to detect any differences between high and low 
use sites for each of the three habitat types. All sites 
designated as low use had significantly fewer (P<0.05) 
roads relative to the high use areas for a given habitat 
type. 
Survey sites are abbreviated to 3-letter designa¬ 
tions: savannah (Sav), riparian (Rip), and upland (Upl) 
with numbers signifying high (1) or low (2) military use 
(Fig. 1). Savannah sites were located in training area 43 
in northwestern Fort Hood (Sav 1) and training area 27 
in south Fort Flood (Sav 2). Upland sites were located in 
training area 44 in north central Fort Hood (Upl 1) and 
training areas 2,3, and 5 in east Fort Hood (Upl 2). Ri¬ 
parian sites were located in training area 53 in north Fort 
Hood (Rip 1) and in training area 8 in east Fort Hood 
(Rip 2). 
Each site was surveyed once per season (begin¬ 
ning in December 1995 through January 1997). Surveys 
were conducted in the following order: Sav 2, Sav 1, Upl 
1, Rip 2, Upl 2, Rip 1. This sequence was maintained to 
ensure that at least five weeks had passed from the time 
of one survey of an area until the next survey of that area 
in the following season. This allowed adequate time for 
captured animals to remix with the population (Begon, 
1979) as well as to ensure that each survey represented a 
true sampling of each season (i.e., an area was not sampled 
at the end of one season then subsequently sampled the 
following week, during the next season). The only viola¬ 
tion of this sampling series occurred during the autumn 
survey. Deer hunting within two of the survey areas re¬ 
quired work in areas where no hunting was occurring. 
One area (Sav 1) was sampled during deer season with 
hunting activity ongoing. 
