The nutritive value of palm kernel meal measured in vivo and using rumen fluid and enzymatic techniques

The objectives of this experiment were to determine the in vivo digestibili ty of solvent extracted and expeller palm kernel meal (PKM) and to establis h how well this was predicted by laboratory techniques. Eight samples of ex peller PKM and four samples of solvent extracted PKM were collected over 4 years. Their in vivo digestibility was measured by total faecal collection using four wethers per sample. The laboratory methods for predicting digest ibility were the in vitro rumen fluid (RF), neutral detergent cellulase wit h gammanase (NCG), and pepsin cellulase with gammanase (PCG) methods. Gamma nase has been added to the latter two methods specifically to improve the p rediction of digestibility of PKM. The solvent extracted samples had higher organic matter digestibility (691 vs. 653 g/kg, sed 15.7) and crude protei n digestibility (727 vs. 597 g/kg, sed 36.6). However, the expeller samples had higher (P = 0.12) contents of digestible energy (13.4 vs. 12.5 MJ/kg D Ml, sed 0.48) due to their higher gross energies (20.6 vs. 19.1 MJ/kg DM). None of the laboratory digestibility tests predicted digestibility satisfac torily. The residual standard deviation was 29, 32 and 31 g/kg for the RF, NCG and PCG methods, respectively. In particular, two expeller samples and two solvent extracted samples were grossly underpredicted by the laboratory methods. These results indicate that PKM is a medium quality energy feed f or ruminants and that enzymatic procedures including gammanase or the in vi tro rumen fluid method do not accurately predict its digestibility.