2.5 Statistical analysis
The experimental design was completely randomised with four treatments and ten replications. Data were analysed using the MIXED procedure (Littell et al., 1996 and SAS, 2004). Animal performance variables (apparent digestibility coefficients of diets, animal performance, feed intake and feed efficiency) were analysed using the glycerine levels as fixed factor. In this analysis, the least squares means method was used in the multiple comparison analysis (SAS, 2004).