The long-term effects of adding chestnut (CHE; Castanea sativa) and valonea (VAL; Quercus valonea) tannin-rich extracts to sheep feed were investigated. In Experiment 1, sheep (65 kg BW) were fed 842 g/day of a ryegrass-based hay. The control-treated animals (CON) received 464 g/day of concentrate, and tannin-treated animals received the same amount of concentrate additionally containing 20 g of the respective tannin-rich extract. Hay and concentrates were offered together in one meal. After the onset of treatment, methane release was measured in respiration chambers for 23.5-h intervals (nine times) in a 190-days period. Faeces and urine were collected three times (including once before the onset of the tannin treatment) to assess digestibility and urinary excretion of purine derivatives. Based on the results obtained from Experiment 1, a second experiment (Experiment 2) was initiated, in which the daily tannin dosage was almost doubled (from 0.9 (Experiment 1) to 1.7 g/kg BW0.75). With the exception of the dosage and duration of the treatment (85 days), Experiment 2 followed the same design as Experiment 1, with the same measurements. In an attempt to compare in vitro and in vivo effects of tannin supplementation, the same substrates and tannin treatments were examined in the Hohenheim gas test. In vitro methane production was not significantly different between treatments. None of the tannin-rich extract doses induced a reduction in methane in the sheep experiments. On the 1st day of tannin feeding in both experiments, tannin inclusion tended to decrease methane release, but this trend disappeared by day 14 in both experiments. In balance period 3 of Experiment 1, lower dry matter and organic matter digestibility was noted for tannin treatments. The digestibility of CP, but not NDF or ADF, was reduced in both experiments. A significant shift in N excretion from urine to faeces was observed for both tannin-rich extracts in both experiments, particularly in Experiment 2. In balance period 2 of Experiment 2, an increased intake of metabolisable energy for VAL was observed. The urinary excretion of purine derivatives was not significantly different between treatments, indicating that microbial protein synthesis was equal for all treatments. Thus, we concluded that both tannin-rich extracts temporary affect processes in the rumen but did not alter methane release over a longer period.