If the data on total increase in test diameter and wet weight, the total
wet and dry food ingested, and the FCR failed to achieve normality and/or
homogeneity of variances and could not be transformed, then a
Kruskal–Wallis non-parametric one-way ANOVA on ranks followed by
Dunn's post hoc test was used to determine differences among the
treatments (Zar, 1996). All data are presented as the mean ± standard
error (SE).