The apparent digestibility of diets containing fish meal, soybean meal or bacterial meal fed to Atlantic salmon (Salmo salar): evaluation of different faecal collection methods

Digestibility measurements obtained from four faecal collection methods wer e compared in two experiments with Atlantic salmon, reared in freshwater. T he fish were fed three diets with different protein sources: a diet with 55 % fish meal (FM); a diet with 35% FM and 31% soybean meal; and a diet with 34% FM and 20% bacterial meal. In Expt. 1, faecal dry matter content and ap parent digestibility coefficients (ADC) of macronutrients were compared usi ng three different methods of faeces collection: stripping, mechanical siev ing of faeces from the outlet water of the tanks and immediate freezing (si eving), and by dissection. In Expt. 2, stripping and sieving were compared to a procedure where the ice and faeces mixture from sieving was freeze dri ed prior to analysis (sieving with FD). Faecal dry matter content, and the ADCs of organic matter and fat differed significantly between all three die ts (soybean meal diet < bacterial meal diet < fish meal diet). ADC of nitro gen (N) was significantly higher in the fish meal diet than in the other tw o diets in Expt. 1, whereas in Expt. 2, ADC of N was equal in the fish meal and soybean meal diets. The soybean meal and bacterial meal diets were ran ked to give lower ADC of energy than the fish meal diet. For faecal dry mat ter content, all three collection methods used in Expt. 1 were ranked diffe rently (stripping < sieving < dissection). For ADC of organic matter and N, sieving produced significantly higher means than the other two methods. In Expt. 2, the same ranking of the three different collection methods used ( stripping < sieving with FD < sieving) was seen for ADC of organic matter, N and energy. Faecal collection method had no effect on the estimates of AD C of fat. Diet differences explained 45% of the variation in ADC of organic matter and 14% for ADC of N in the two experiments. Faecal collection meth od explained 42% of the variation in ADC of organic matter and 63% of the v ariation in ADC of N. Small, but significant, interactions between diet and collection method were seen for these two digestibility estimates. The res ults obtained using the different faecal collection methods were highly cor related within experiment for all estimates with the exception of the ADC o f N in Expt. 2. No significant differences were seen between the two experi ments with respect to ADC of organic matter and N, and the estimates from t he two experiments were highly correlated. (C) 1998 Elsevier Science B.V. A ll rights reserved.