Methane emissions changed nonlinearly with graded substitution of alfalfa silage with corn silage and corn grain in the diet of sheep and relation with rumen fermentation characteristics in vivo and in vitro1,2

Feeding grain and corn silage have been proposed as practices to reduce enteric methane (CH) emissions per unit of intake from ruminants, but the inclusion level required in the diet is normally not specified. The objectives of the current study were to determine the CH emission factor (g/kg DMI) of sheep fed alfalfa silage substituted with increasing levels of corn silage or corn grain at a fixed DMI level (2% of BW) and determine its relationship with rumen fermentation characteristics and microbial community composition and with in vitro fermentation characteristics of the same diets incubated using a standard laboratory method. Romney ewe hoggets (approximately 14 mo old; = 64) were randomly allocated to 8 dietary treatments, which included chaffed alfalfa silage alone or substituted with either 25, 50, 75 or 100% corn silage or 25, 50 or 65% rolled corn grain on a DM basis. After acclimatization to the diet, DMI and CH emissions were measured from individual sheep for 2 consecutive days in open-circuit respiration chambers and a rumen sample was collected at 3 h after feeding. The same diets were also incubated in an automated in vitro gas production system for 48 h using rumen liquid of fistulated nonlactating dairy cows grazing pasture. Increasing the substitution of alfalfa silage with corn silage or corn grain in the diet of sheep resulted in a quadratic response ( < 0.01) in CH emissions per unit of DMI (CH/DMI) with either supplement. For both supplements, CH/DMI increased in mixtures of up to 50% supplement inclusion and then decreased with greater supplement inclusion, especially with corn grain inclusion, but the level did not fall below that for 100% alfalfa silage. The ratio of acetate + butyrate to propionate + valerate and the propionate proportion alone in rumen liquid were the strongest single predictors for CH/DMI in the overall data set and explained 37.1 and 32.5%, respectively, of the variation in CH/DMI. Methanogens of (21.1% of total methanogens; = 0.247) and (10.7% of total methanogens; = -0.411) clades had weak to moderate correlations with in vivo CH/DMI. There was a weak quadratic relationship ( < 0.35) between in vivo CH/DMI and the in vitro parameters of gas and CH production and total VFA, whereas there was a moderate relationship ( = -0.50) between in vivo CH/estimated rumen degradable carbohydrates and in vitro CH/DM. In conclusion, CH/DMI changed in a nonlinear fashion with increasing supplement inclusion in the alfalfa forage diet when fed at 2% of BW to sheep; however, implications on predicting its influence on greenhouse gas emissions per unit of animal product, for whole farm emissions in life cycle analysis or total national emissions in the national inventories, should be determined.