In each of four experiments, forages (Experiment 1, 20 hays from Napier and Pigeon grass cut at different intervals; Experiment 2, eight leaves and branches from tree lucerne harvested at different intervals; Experiment 3, six hays from Reed Canary grass cut at different intervals; Experiment 4, 15 crop residues from sorghum, tef, wheat and barley) were assessed for in situ rumen dry matter (DM) disappearance, in vitro gas production using a pressure transducer technique (PTT), chemical composition (nitrogen and fibre fractions) and in vitro digestibility (DMD). In situ DM disappearance data were fitted to the model of Ørskov and McDonald, i.e. P = a o + b o(1 − exp(− c o t)), where P is the DM disappearance at time t, a o and b o are fitted constants and c o (h −1) is the fractional rate of DM disappearance. PTT data were fitted to the model of France et al., i.e. Y=A(1 − e −b(1−T)−c( t − T ) ) where Y is gas production after time t, b and c are rate constants, A is the asymptote and T is lag time. In situ DM disappearance parameters were linearly regressed against in vitro gas production parameters, chemical composition and in vitro DMD. Overall c o of forages was correlated ( P < 0.01) with fractional rate of gas production μ ( μ = b + c 2√t ). c o was also correlated with nitrogen, fibre fractions and in vitro DMD. In all experiments, effective in situ DM disappearance (EDMD: DM degraded in the rumen at a given rate of passage) was significantly ( P < 0.01) correlated with extent of gas production (at the assumed passage rate from the rumen). EDMD was also correlated ( P < 0.01) with in vitro DMD in all experiments, and with chemical composition (N, NDF, ADF, KMnO 4 lignin). It is concluded that in situ DM disappearance parameters of forages such as used in the present study, may be predicted from in vitro gas production parameters, and to some extent, also from chemical composition and in vitro digestibility. However, the accuracy of prediction is likely to vary, depending on the type of forages used.