Pereira-da-Silva et al., 2000. Rev. Bras. Zootec., 29 (5): 1273-1280

The response of Nile tilapia to the attraction and palatability of 14 ingredients was evaluated. The two-choice method was used, comparing each pelleted ingredient with a control diet. Four aquaria (750 litres) were used each containing three fingerlings and two feed containers at the right and left corners. The time elapsing between giving the feed ingredient and arrival of fingerlings to the feed containers was measured, as well as visiting frequency, amount of ingested pellets and ejection frequency. Treatments were compared by the non-parametric method of Kruskal-Wallis and the comparison of ingredients for the variable set using grouping analysis and principal components analysis. The number of ingested pellets (in the morning and in the afternoon) and the visiting frequency to the feed containers (afternoon) were considered to be the most discriminatory variables. The ingredients were classified as follows: (a) low attraction and palatability = wheat meal, soyabean meal, cottonseed meal, cassava meal, cassava scrapings, sunflower meal and corn [maize] meal, (b) medium attraction and palatability = sugarcane yeast and corn gluten meal and (c) high attraction and palatability = integral lyophilized egg, silk worm meal, fish meal, meat meal, shrimp meal. The frequency of ejected pellets and time spent in approaching feed (in morning and afternoon) were considered to be the least discriminatory variables. The behaviour response of fishes changed according to the presented ingredient.