The bacterial counts were log10-transformed prior to
analysis to ensure normality of data. Statistical analyses of
growth performance were similar to that in experiment 1.
The effect of weaning age(3,4and5weeks)was tested in
a randomised complete block design with gender,ancestry
and weaning weight as blocking factors. The pair mean
was the experimental unit for growth performance analyses.For bacterial counts,the individual pig was the
experimental unit.The results were presented as least
square meansSEM. Tukey Studentized Range(HSD)test
was used for means separation.The probability level that
denotes significance was P>0.05.