and Levene's test respectively. Non-normal and heterogeneous datawere
transformed until normality and homogeneity were achieved.
Two-way analysis of variance (ANOVA)was used to compare the effects
of treatments on the body weight, SGR, IR, FPR, FCE and energy
budget, followed by Tukey multiple comparison post hoc test (Zar,
1999). Differences were considered significant at a probability level of
0.05. Before multiple comparisons, the interaction of the two factors
was analyzed. For the data with a non-significant interaction (P N 0.05),
the interaction item was eliminated from the ANOVA model and the
main effects of treatments were further compared. As for the data with
significant interaction (P b 0.05), the effects of each treatment were
analyzed within the respective level of another treatment with Bonferroni
adjustment to correct the magnification of type I error (Corston and