Irrigated rice fields are an important source of atmospheric methane. In order to improve our understanding of the controlling processes, we measured in situ CH4 emission and CH4 oxidation in an Italian rice field in 1998 and 1999, and studied CH4 production in soil and root samples. The CH4 emission rates were correlated with diurnal temperature variations and showed pronounced seasonal and interannual variations. The contribution of CH4 oxidation to total CH4 flux, determined by specific inhibition with difluoromethane, decreased from 40% at the beginning to zero at the end of the season. The stable carbon isotopic composition of the emitted CH4 also decreased. The CH4-oxidizing bacteria probably became limited by nitrogen as indicated by the seasonal decrease of NH4+. Thus, CH4 oxidation had little effect on CH4 emission. Methane production on rice roots was relatively constant over the season. Methane production in soil slowly increased after flooding and was highest in the middle of the season. Pore water concentrations of CH4 showed a similar seasonal pattern. In 1999, CH4 production increased later in the season and reached lower rates than in 1998. An additional drainage in 1999 resulted in higher ferric iron concentrations, higher soil redox potentials and lower acetate concentrations. As a result, acetate-utilizing methanogens were probably out-competed by iron-reducers so that a larger percentage of [2–14C]acetate was converted to 14CO2 instead of 14CH4. The residual CH4 production was relatively low and was mainly due to H2/CO2-dependent methanogenesis. Experiments with radioactive bicarbonate and with methyl fluoride as specific inhibitor showed that the theoretical ratio of 7:3 of methanogenesis from acetate vs. H2/CO2 was only reached later in the season when total CH4 production was at the maximum. In conclusion, our results give a mechanistic explanation for the intraseasonal and interannual differences in CH4 emission.