Wild Hog Rooting 
171 
10-m spacing. Trapping was conducted from 3 July through 11 Au- 
gust in 1989 and 28 June through 4 August in 1990. We baited 
traps with sunflower seeds packaged in a bag of cheesecloth and 
suspended in the back of the cage. Baiting traps with a mixture of 
rolled oats and peanut butter in the beginning of 1989 attracted 
foraging bears ( Ursus americanus ) to the trap sites. These bears 
destroyed a significant number of traps at each site, effectively shut- 
ting down the trap grid. Each captured small mammal was identi- 
fied to species, sexed, aged (juvenile or adult), weighed, toe clipped 
for future identification, and released. 
In 1990, we added pit-fall traps because of the lack of shrew 
captures during the 1989 field season. Pits were constructed by 
burying a 1-gallon plastic planter along a fallen log, lining the 
planter with plastic, and filling it about one-third full with water. 
Ten pit-fall traps were interspersed among each live trap site, with 
one or two pit-fall traps per live trap line. All animals collected 
from pit-falls were identified according to species, weighed, and 
when possible, aged and sexed. 
Data Analysis — We used Program CAPTURE (Otis et. al. 1978, 
White et. al. 1982) to analyze data from the Sherman live traps for 
the capture of deer mice. CAPTURE selected the appropriate popu- 
lation model for the data and calculated a population estimate and 
capture probabilities for each site. Because of insufficient capture 
rates, no data on small mammal species other than deer mice from 
the live traps and no data from the pit-fall traps were analyzed 
with CAPTURE. We combined the results of the 1989 and 1990 
CAPTURE analyses and tested them with a two-way ANOVA (al- 
pha level = 0.05) to determine if (1) pooling the data would reveal 
significant differences between rooted and unrooted sites, (2) there 
were differences in overall populations between the two years, and 
(3) there were any signifiant interaction effects between year and 
rooting class. 
Vegetation 
Sampling — We selected 20 variable classes for measurement at 
each trap. Table 1 contains a list of the variable classes measured 
and the method by which measurements were taken. Several of the 
classes (e.g., percent species cover) involved measurements for each 
plant species represented. Therefore, the number of variables consid- 
ered is greater than the number of variable classes. The point-quar- 
ter method for sampling trees followed that suggested by Phillips 
(1959). We defined overstory as woody vegetation >7.5-cm dbh and 
