Direct-fed microbials (DFM) are used to modulate ruminal function and induce beneficial effects on ruminants. The objectives of this work were to 1) screen bacterial strains for their antimethanogenic potential in vitro and 2) assess the effect of 3 selected DFM on ruminal methane (CH) emissions, fermentation parameters, and microbial profiles in sheep. Forty-five bacterial strains were preselected based on their metabolism and fermentation characteristics. These bacteria were screened for their ability to reduce ruminal methanogenesis using 24-h batch incubations and an inoculum of 10 cfu/mL of medium. The addition of bacterial strains stimulated ruminal fermentation with increases in total gas production for 41 strains ( < 0.05) without a concomitant increase in CH production (only 9 strains had higher CH than the controls without DFM; < 0.05). 53-W, D31, and D1 had the greatest difference between total gas and CH production and were selected for further in vivo testing. Twelve rumen-cannulated Texel wethers were divided into 3 groups and were treated daily for 4 wk with 6 × 10 cfu/animal for and and 3 × 10 cfu/animal for . Measures of enteric CH, ruminal fermentation, and ruminal microbial traits were performed before, at 2 and 4 wk during the treatment period, and at 2 wk after the DFM treatment stopped. Methane production was reduced by 13% ( < 0.05) with after 2 wk of DFM administration, and this effect was maintained throughout the treatment and posttreatment periods. In contrast, had no effect on CH production, and increased it by 16% ( < 0.05) after 4 wk of DFM administration. There was no effect on other fermentation parameters or on the bacterial, archaeal, and protozoal numbers monitored by quantitative PCR. However, denaturing gradient gel electrophoresis profiles indicated changes in bacterial and archaeal diversity in the and groups. Although added bacteria were unable to permanently colonize the rumen, had a greater 24-h survival rate than the others, implying that the persistence of DFM may be important for modulating ruminal traits of interest. These results suggest that bacterial DFM used in this trial were able to modify CH emissions, although correlated changes in other ruminal parameters studied were minor.