Prediction of digestible energy and gross energy digestibility of feeds and diets in ostriches.

1. The aim of this study was to assess regression equations able to predict the digestible energy (DE) and gross energy digestibility (GEd) of feed ingredients and diets for ostriches. 2. Results of chemical-nutritional characteristics from 17 ingredients (two varieties of maize, two barleys, oat, triticale, wheat bran, soybean meal, sunflower meal, beet pulp, maize silage, alfalfa hay, 4 alfalfa meals and lupin) and 12 experimental diets were used in a stepwise procedure. 3. Acid detergent lignin (ADL) was the first independent variable included in the model to predict the DE of all the samples (R(2) = 0.65 and Residual Standard Deviation (RSD) 1.02). When the concentration of ash, acid detergent fibre (ADF) and crude fibre were included in the model, the R(2) value of the regression equation increased (from 0.65 to 0.85) and RSD decreased (from 1.02 to 0.48). 4. The ADL concentration was also the first independent variable chosen by the stepwise regression analysis for the estimation of GEd from chemical-nutritional characteristics of feeds, explaining 57% of the total GEd variation. The concentrations of crude protein and ADF included at the second and third steps in the model increased the R(2) (up to 0.70 and 0.73, respectively) and decreased the RSD values (from 0.29 to 0.21 and 0.20, respectively). When other variables as crude fibre, ash and gross energy were included in the model, the coefficient of determination and the RSD strongly improved (0.85 and 0.12, respectively).