FACTORS AFFECTING THE INTAKE OF GRASS-SILAGE BY CATTLE AND PREDICTIONOF SILAGE INTAKE

A partially balanced change-over design experiment involving 192 beef steers, which were initially 14 months old and 415 kg live weight, was carried out to determine the intakes of 136 silages from commercial f arms in Northern Ireland. Each silage was offered ad libitum as the so le food to 10 animals, with eight silages offered in each of 17 period s over 2 years. A standard grass hay was offered to 16 animals in each period to enable period effects on intake to be removed. Detailed che mical and biological compositions of the silages were also determined. The ranges for pH and dry matter (DM), crude protein, ammonia-nitroge n and apparent digestible organic matter (in vivo) concentrations in t he silages and silage dry DM intakes were 3.50 to 5.49 (s.d. 0.396); 1 55 to 413 (s.d. 4.31) g/kg; 79 to 212 (s.d. 24.4) g/kg DM; 45 to 384 ( s.d. 63.2) g/kg total nitrogen; 528 to 769 (s.d. 58) g/kg DM and 4.3 t o 10.9 (s.d. 1.13) kg/day respectively. Relationships between intake a nd individual parameters or groups of parameters have been developed u sing simple and multiple linear regression analysis and partial least- squares analyses. Silage intake was closely related to factors which i nfluence the extent of digestion and rate of passage of the material t hrough the animal, as indicated by the strong relationships (R-2 of re gressions = 0.28 to 0.50) with in vivo apparent digestibility and rume n degradability and the concentrations of the fibre and nitrogen facto rs. Intake was poorly correlated with factors such as pH, total acidit y, buffering capacity and the concentrations of lactic, acetic and but yric acids (R-2 of regressions = zero to 0.11). Near infrared reflecta nce spectrometry (NIRS) provided the best fit relationship with intake (R-2 of relationship = 0 90). The results also indicate that the inta ke potential of silages can be directly predicted with a high degree o f accuracy from the NIRS of both dried and undried samples of silage, provided the appropriate sample preparation and scanning methods are u sed.