Investigation of the effects of substitution levels, assay methods and length of adaptation to experimental diets on determined metabolisable energy value of maize, barley and soya bean meal

1. Two experiments were carried out to investigate the influence of substitution levels, assay methods and adaptation length of broilers to experimental diets on metabolisable energy of corn and barley (Experiment 1) and soya bean meal (Experiment 2). In each experiment, the test feedstuffs proportionately replaced the reference diet at the rates of 150, 300 or 450 g/kg. Each treatment had 8 or 7 replicate cages containing three birds per cage. Broiler chickens were adapted to experimental diets for either 7 or 4 d prior to excreta collection on d 20 and 21.2. Metabolisable energy was determined using the difference method at each substitution level and the regression method, using the 4 combinations of the substitution levels.3. In Experiment 1, AME and AMEn were lower (P < 0.01) for barley than for maize, irrespective of the assay method and there was a trend for an adaptation × cereal grains interaction (P < 0.10) for AME determined by the difference regression method. For maize, AME and AMEn calculated using regression or the difference method were not significantly different, whereas barley AME and AMEn values were greater (P < 0.01) when derived using regression analysis.4. In Experiment 2, there was no significant adaptation length × substitution levels interaction. The AME and AMEn values were lowest (P < 0.01) at 150 g/kg SBM inclusion level when calculated using the difference method. On the other hand, AME and AMEn values determined by the regression method were not influenced by substitution levels.5. It was concluded that the influence of substitution levels on assayed metabolisable energy is feedstuff-dependent and that the regression method produced more consistent metabolisable energy values in a feedstuff-independent manner.