The gas dilution technique was used to evaluate the possibility of estimating the volume of gaseous phase in the rumen from its composition in sheep given rice whole crop silage (RWS) or dent corn silage (DCS) at a level of maintenance (M) or 2 M, and in the course of fasting. The rumen gas composition was determined at 2 and 7.5 h after morning feeding. Nitrogen gas was injected by using an airtight syringe into the rumen immediately after collecting the rumen gas sample as a control. Then rumen gas samples were collected at 5, 10, 20, 40 and 60 min. after injection. Dry-matter intakes were
for DCS, and
for RWS, at 1 M and 2 M levels, respectively. Animals ingested both silages about 20% less than expected at 2 M level. The rumen gas composition did not differ significantly between 2 h and 7.5 h after feeding except for
. Content of
in gas composition was significantly higher at 2 M level than at 1 M (p<0.05) for both RWS and DCS, whereas
showed no significant difference between feeding levels. At both feeding levels,
showed a higher (p<0.05) percentage in DCS than RWS. A dilution technique by using
injection is not appropriate for the determination of gas production in vivo, unless the rate of rumen gas turnover is considered. Changes in composition at fasting indicate that the rumen fermentation may reach the lowest level after 72 h fasting for sheep given silage as their sole diet.