2.6. Statistical analysis
In the current study, all data were subjected to statistical analysis in a randomized complete block design using the GLM procedures (SAS Inst. Inc., Cary, NC), with the pen serving as the experimental unit. The initial BW was used as a covariate for ADFI and ADG, and initial values were used as a
covariate for the blood profile. Duncan's multiple range test was used to compare the means of the treatments. Variability in the data was expressed as the pooled standard error (SE) and a Pb0.05 was considered to be statistically significant.