Effect of maize silage to grass silage ratio and feed particle size on ruminal fermentation in vitro

The effect of the forage source on ruminal fermentation in vitro was investigated for fine (F) and coarse (C) milled diets, using a modified Hohenheim gas production test and a semi-continuous rumen simulation technique (Rusitec). It was hypothesised that the replacement of maize silage by grass silage might lead to associative effects and that interactions related to particle size variation could occur. Five diets with a maize silage to grass silage ratio of 100 : 0, 79 : 21, 52 : 48, 24 : 76 and 0 : 100 differed in their content of CP and carbohydrate fractions, as well as digestible crude nutrients, derived from a digestibility trial with wether sheep. For in vitro investigations, the diets were ground to pass a sieve of either 1 mm (F) or 4 mm (C) perforation. Cumulative gas production was recorded during 93 h of incubation and its capacity decreased with increasing proportion of grass silage in the diet. Across all diets, gas production was delayed in C treatments compared with F treatments. Degradation of crude nutrients and detergent fibre fractions was determined in a Rusitec system. Daily amounts of NH(3)-N and short-chain fatty acids (SCFA) were measured in the effluent. Degradation of organic matter (OM) and fibre fractions, as well as amounts of NH3-N, increased with stepwise replacement of maize silage by grass silage. Degradability of CP was unaffected by diet composition, as well as total SCFA production. In contrast to the results of the gas production test, degradation of OM and CP was higher in C than in F treatments, accompanied by higher amounts of NH3-N and SCFA. Interactions of silage ratio and particle size were rare. It was concluded that the stepwise replacement of maize silage by grass silage might lead to a linear response of most fermentation characteristics in vitro. This linear effect was also supported by total tract digestibility data. However, further investigations with silages of variable quality seem to be necessary.