The calculations used are summarized in Table 3. Dry matterbelow 1.2 m and dissolved material was estimated indirectly inthe sieving experiments by subtracting the sum of recovered drymatter from total dry matter supplied. Normality (p > 0.05) andhomogeneity of variance (p > 0.05) of the data was verified using theShapiro–Wilk test and the Levene’s test, respectively. Fecal wastecomposition was compared using a t-test, assuming unequal vari-ances (n = 2). The effect of diet on particulate dry matter content, drymatter median, particle size median, microparticle production andinitial COD degradability was analyzed using one-way ANOVA. Theeffect of sonication time and diet composition on particle size distri-bution was analyzed using a two-way ANOVA, testing for an effectof diet (D), sonication time (T), and an interaction effect (D × T).The data required log10-transformation before analysis with a two-way ANOVA