CH4 release rates from rice paddies were measured in Vercelli, Italy, in 1983 during a complete vegetation period, using a static box system. The rice paddies were found to be a source of atmospheric methane during the time of flooding. The CH4 release rates range between a few milligrams of CH4 per square meter per hour and 51 mg CH4 m−2 h−1 and show a seasonal variation with maximum emission rates between tillering and flowering. Averaged over the complete vegetation period, the CH4 release rate accounts for 16 mg m−2 h−1. Significant differences in the CH4 release rates of unfertilized field plots and plots fertilized with mineral fertilizer (CaCN2) were not observed. The CH4 release rates show strong diurnal variations, with highest values in the late afternoon and lowest values in the early morning, which coincides with the temperature variation in the upper soil layer (1–10 cm). These variations must be taken into account in estimating reliable global CH4 emission rates from rice paddies. CH4 is almost exclusively emitted into the atmosphere by gas bubbles during the first 6 weeks after flooding the rice paddies, that is, on fields without vegetation. Afterwards, 80% of the observed CH4 transport from the paddy soil into the atmosphere was due to diffusion through the stems of the rice plants. Transport by diffusion through the paddy water appeared to be of minor importance. The global annual CH4 emission from rice paddies is estimated to be of the order of 39–94 Tg yr−1 (Tg = 1012 g) for 1940 conditions and 70–170 Tg yr−1 for 1979 conditions, indicating a secular trend of the CH4 emission by rice paddies of about 1.6% per year during the last 35 years.