The two experimental feeds were formulated to be isoener-getic and isonitrogenous on digestible energy and nitrogen, andto produce feces with a contrast in fiber content between the treat-ments. Rainbow trout (Oncorhynchus mykiss, Walbaum 1792) werefed a low starch, high NSP (HNSP) diet and a high starch, low NSP(LNSP) diet (Table 1). We used extruded diets, which were producedby Research Diet Services (Wijk bij Duursteede, The Netherlands).Acid insoluble ash was used as a marker to determine digestibil-ity (Atkinson et al., 1984; Vandenberg and De La Noue, 2001). Thefish were fed a floating pellet of 3 mm, and the daily ration (∼1.5%body weight/d over the whole experimental period) was equallydivided between morning (9:30 h) and afternoon feeding (16:30 h).Feeding was done by hand and under close observation to avoidleftover pellets in the feces collection. In case a tank would noteat the whole ration, the leftover pellets would be collected andcounted. In the subsequent feeding, the ration for the remainingtanks would be reduced by the amount of uneaten feed recovered toensure homogenous growth between tanks and diets. Feed sampleswere collected weekly and analyzed as a pooled sample at the endof the experiment to determine proximate composition