A meta-analysis was conducted to evaluate the effects of protozoa concentration on methane emission from ruminants. A database was built from 59 publications reporting data from 76 in vivo experiments. The experiments included in the database recorded methane production and rumen protozoa concentration measured on the same groups of animals. Quantitative data such as diet chemical composition, rumen fermentation and microbial parameters, and qualitative information such as methane mitigation strategies were also collected. In the database, 31% of the experiments reported a concomitant reduction of both protozoa concentration and methane emission (g/kg dry matter intake). Nearly all of these experiments tested lipids as methane mitigation strategies. By contrast, 21% of the experiments reported a variation in methane emission without changes in protozoa numbers, indicating that methanogenesis is also regulated by other mechanisms not involving protozoa. Experiments that used chemical compounds as an antimethanogenic treatment belonged to this group. The relationship between methane emission and protozoa concentration was studied with a variance−covariance model, with experiment as a fixed effect. The experiments included in the analysis had a within-experiment variation of protozoa concentration higher than 5.3 log10 cells/ml corresponding to the average s.e.m. of the database for this variable. To detect potential interfering factors for the relationship, the influence of several qualitative and quantitative secondary factors was tested. This meta-analysis showed a significant linear relationship between methane emission and protozoa concentration: methane (g/kg dry matter intake)=−30.7+8.14×protozoa (log10 cells/ml) with 28 experiments (91 treatments), residual mean square error=1.94 and adjusted R 2=0.90. The proportion of butyrate in the rumen positively influenced the least square means of this relationship