แปลบทความวิจัยAll data were analysed using PROC MIXED (Version 9.2; SAS Inst. Inc., Cary, NC) with diet as fixed effect and block asrandom effect in the model. Pen was the experimental unit. Nutrient digestibility and DE and calculated NE values wereanalysed without using initial BW as covariate. Growth performance data were analysed as repeated measures using weeklypen data with the unstructured (UN) covariance structure based on −2 Res Log Likelihood fit statistics and initial BW as acovariate if significant. Single degree of freedom orthogonal contrasts were used to determine linear or quadratic effects ofwheat millrun inclusion on digestibility and performance for each week or the entire trial (Littell et al., 2006). To test thehypotheses, P < 0.05 was considered significant.