A meta-analysis was conducted to evaluate the extent to which dietary tannin level is related to methane emissions from ruminants. Data from a total of 30 experiments comprising 171 treatments were entered in a database. In vitro batch culture and in vivo measurements were distinguished as experimental approaches. With any approach, methane declined when dietary tannins increased. The in vitro approach predicted the in vivo response quite accurately. However, in vitro, the response followed a quadratic response pattern (R2 = 0.66; lower response with increasing tannin level), whereas in vivo, this decline was linear (R2 = 0.29). This indicates that the in vitro batch culture is of limited accuracy for estimating effects at levels >100 g tannin/kg dry matter. The large variation in methane/digestible organic matter (OM) found at low tannin levels may explain contrasting literature reports. Methane reduction with tannins was associated with a reduced apparent digestion of OM, and especially fibre, but methane/apparently digestible OM declined also. The present findings are helpful as they identified an underlying general antimethanogenic effect of tannins across tannin sources and experimental conditions, thus allowing concentrating the search on sources with satisfactory palatability and low adverse effects on animal performance.