Statistical Analysis
Data for all experiments were analyzed by ANOVA
as randomized complete block design using the GLM
procedures of SAS (SAS Institute Inc., Cary, NC) as
outlined by Steel and Torrie (1980). In EXP 1, treatments
were replicated in 12 blocks with 4 treatments
randomized in each block. In EXP 2, treatments were
replicated in 8 blocks with 6 treatments randomized
in each block. Each pen of broilers served as the experimental
unit in both experiments. Treatment differences
were considered significant with an alpha =
0.05. The PDIFF option of SAS was used to compare
individual diets to the control diet. Orthogonalcontrasts were used to determine significant linear and
quadratic effects of the encapsulated source of butyric
acid on response variables. The response variables for
performance included BW gain, feed consumption, feed
conversion, and mortality. Intestinal morphology data
were analyzed as a one-way ANOVA by the GLM procedure
of SAS. The statistical model included the main
effect of dietary treatment, but BW of selected chickens
was included as a covariate.