This experiment compared linear relationships among end-products of rumen
fermentation measured at the time (t½) at which a feed produces half of its asymptotic gas production) or at 48 h. Meadow hay and corn grain were incubated for t½ (16 and 9 h, respectively) or for 48 h into glass bottles. Each bottle (310 ml) was filled with feed sample (0.5 g) and 75 ml of buffered rumen fluid, and incubated at 39.0°C. Gas production (GP) was measured using the ANKOMRF System, and gas accumulated in headspace of bottles was released at 3.4 kPa. At t½ or 48 h, fermentation fluids were analysed for ammonia N (N-NH3), volatile fatty acids (VFA), residual NDF and N bound to residual NDF (N-NDF). Values of GP were also predicted from VFA. Microbial N (MN) was computed as the difference between N present at the beginning and at the end of incubation. At 48 h, the relationship between GP measured and predicted from VFA was weak (R2 = 0.67; equation not shown), whereas the linear relationship was better at t½ (R2 = 0.94). At t½, the relationship
between N-NH3 and measured GP was strong (R2 = 0.84), as well as that between MN and measured GP (R2 = 0.92).Conversely, these variables were not well related at 48 h. At t½, the valerate content in rumen fl uid was negligible. However, relatively large amounts of valerate were measured after 48 h, probably the result of microbial lysis. Results suggest that relationships among end-products of rumen fermentation can be more accurately evaluated at a substrate-specifi c incubation time (t½) rather than at 48 h.