analysisAll data on diet-associated difference were analyzed by GLM procedure appropriate for completely randomized design(SAS, 2002). The pen was an experimental unit for performance data analysis while individual birds were the experimentalunit for bacterial and morphological measurements. The orthogonal polynomial contrast test was performed to determinelinear and quadratic effects of increasing inclusion level of MP in diets on each measurement. Broken-line regression analysiswas employed using NLIN procedure (Robbins et al., 2006).