evaluated for that study). Presumably, expressing GPX activity and AP per mg of protein reflects

the proportion of total proteins functioning as antioxidants, whereas expressing these parameters

per [g of DNA reflects antioxidant capacity per cell.

 Data were tested for normality (Shapiro-Wilk test) and homogeneity of variances

(Levene's test) prior to parametric analysis. If either test yielded a significant result (p < 0.05),

data were transformed using a natural log, reciprocal, square root, square, reciprocal square, or

reciprocal square root transformation. If transformation did not improve homoscedasticity and

Tamhane's T2 post hoc test could not be used, data were tested for statistical significance using a

Kruskal-Wallis test. Otherwise, data within each sampling period were tested for statistical

significance using analysis of variance (ANOVA). When statistically significant differences

among t12 treatment groups were found, pairwise comparisons were evaluated using Tukey's

Honestly Significant Difference (HSD) post hoc test (if variances were equal) or Tamhane's T2

post hoc test (if variances were not equal).

 Data were analyzed using SPSS for Windows (Release 11.0.0). For all reported analyses,

data are expressed as means + standard errors (unless otherwise noted) with alpha set at 0.05.

 Results

 Turtles in the AL group grew significantly faster than those in the R group during each

week of the study. After the switch from a restricted to an ad libitum diet (at the end of five

weeks), R-AL turtles grew significantly faster than those in the AL group during weeks 7

through 9. This period of growth compensation ceased prior to the end of the study such that AL

and R-AL turtles were growing at comparable rates by the time t12 samples were collected (as

reported in Chapter 2). Different growth rates yielded significantly different body masses for

turtles in each treatment sampled at the conclusion of week five and at the conclusion of week

twelve (Fig. 4-1 and Table 4-1).