This experiment was carried out to evaluate the fractions that compose carbohydrates and protein from elephant grass silage with 15% of coffee hulls, cocoa meal or cassava meal. The addition of residues at the ensilage moment was performed in natural matter basis (weight/weight), with ten replicates per treatment. The addition of residues at the ensilage moment was performed in natural matter basis (weight/weight), with ten repetitions per treatment. The total carbohydrates level of the silage with cassava meal was greater than that of other silages, with the lowest value observed to silage with cocoa meal, followed by control and silages with coffee hulls. Higher values of no-fiber carbohydrates (A + B1) were also verified for silages with cassava meal. Smaller values of the indigestible fraction (C, %TC) were observed for silage with cassava meal, while silages with coffee hulls and cocoa meal showed higher values for this fraction. Higher CP values were verified in silage with cocoa meal, followed by the silages with coffee hulls, control and with cassava meal. The addition of by-products influenced the protein fractions, with higher values of fraction A (%CP) in control silages and with cassava meal. Although with higher protein value, the higher protein indigestible fraction (C, %CP) values were verified for silages with cocoa meal and coffee hulls. The addition of cocoa meal and coffee hulls to elephant grass at the ensilage moment produced silages with higher carbohydrate and protein indigestible fractions. The cassava meal ensiled with elephant grass showed to be a good addictive to silage production.
by-product; forage conservation; Pennisetum purpureum; residue
