color and pH main effects and interactions were treated as fixed effects and considered significant at pb0.05. For sensory data, panelists were considered repeated measures.
Separation of least square means was achieved using probability of difference, adjusted with the Bonferroni procedure for multiple comparisons.