LAUSR

Forty samples of grass silages, made from the five most widely used grass species in rumi - nant nutrition ( Dactylis glomerata L., Phleum pratense L., Lolium perenne L., Festuca arundinacea L. and hybrid Felina) were tested in the present study. These grass species conserved by an ensiling process were compared among each other according to dry matter (DM) rumen degradability parameters ( a = portion of DM solubilized at initiation of incubation, b = fraction of DM potentially degradable in the rumen, c = rate constant of disappearance of fraction b and ED DM = effective degradability of DM, estimated for each ingredient assuming the rumen solid outflow rates of 0.02 (ED DM2 ), 0.05 (ED DM5 ) and 0.08 (ED DM8 ) h -1 ). Based on the chemical composition of grass silages the regression equations for prediction of ED DM were evaluated. The influence of the ensiling process on dry matter degradability parameters was also assessed. The best values of ED DM were determined for Lolium perenne (ED DM2 = 753.2, ED DM5 = 631.1 and ED DM8 = 567.7 g/kg DM). The best predictor was NDF ( R 2 -values of 0.757 (ED DM2 ), 0.863 (ED DM5 ) and 0.906 (ED DM8 )). Using two predictors the accuracy level increased. The combination of CF and NDF gave R 2 -values 0.892, 0.920 and 0.929 for ED DM2 , ED DM5 and ED DM8 , respectively. The regression equations based on the most important grass species harvested in different vegetation periods seem to be a useful tool for practical use. No significant ( P < 0.05) effect of the ensiling process in relation to dry matter rumen degradability param - eters was proved.