The experimental unit was a hen kept in an individual cage so that each treatment group consisted of 48 replications. Although daily and weekly data were recorded, statistical data were analyzed and presented as the mean of each 4-wk period. The PROC GLM statement of SAS Institute Inc. (2006) used a completely randomized design for all 5 variables: ADFI, hen day egg production (HDEP), egg weight, feed per kilogram of eggs, and feed per dozen eggs. An underlying assumption in an ANOVA (using PROC GLM) is that the observational errors are the same for each level of a factor (in our case, for each of the 3 diets: high, medium, and low). Thus, a homogeneity of variance test was run on each variable to assess this.