SMITH: GOPHERUSPOLYPHEMUS ON THE ORDWAY PRESERVE


 Data Analysis


 The sex ratio of tortoises captured over the six-year period was evaluated
using X2 analysis (Caughley 1977). A two sample t-test was used to compare mean
carapace length of male and female tortoises. Recapture rate for each year of the
study was calculated as the proportion of marked animals captured in a year to the
total number of marked tortoises. Mean annual growth increments were calculated
for individuals recaptured one year or more after they had been marked. Linear
regression was used to evaluate the relationship between burrow width and
carapace length, and to derive a least squares regression line for estimating the
carapace length of a tortoise based on burrow width.
 Home range of the 14 radio-instrumented tortoises was calculated using the
minimum convex polygon method (Mohr 1947). Areas were determined using a
digitizer. Analysis of covariance was used to compare mean home range of females
in sandhill and old field habitats after adjusting for carapace length and length of
time followed. A separate variance t-test was used to compare the mean feeding
radius of females in the two habitats.
 Only active and inactive burrows were included in the analysis of population
structure and burrow density. As yet, a site-specific, tortoise-to-burrow correction
factor has not been determined for the Ordway Preserve. Therefore, each active and
inactive burrow was considered to represent a single tortoise although this will
probably overestimate density by a factor of 1.5 to 2. Burrow density, reported in
burrows/ha, was calculated for each location. Burrow width was converted to
estimated carapace length using the regression equation derived from the carapace
length and burrow width measurements. A two-sample t-test and separate variance
t-test were used in the following comparisons of mean estimated carapace length:
core vs. perimeter old field; core vs. perimeter sandhill; core sandhill vs. core old
field; and perimeter sandhill vs. perimeter old field.
 The relationship between female carapace length and clutch size was
 examined using regression analysis. Mean clutch size between years was compared
 using a separate variance t-test (Ott 1988). Two-sample t-tests were used to
 compare nest depth and distance from the burrow entrance between years.
 Hatching success for each year was calculated as the number of eggs that hatched
 successfully divided by the total number of eggs. Hatching success between years
 was compared using )2 analysis (Caughley 1977). Mean carapace length and body
 mass of hatchlings in 1990 and 1991 were compared using separate variance
 t-tests.