Methane emissions from rice fields are affected by a number of environmental and agricultural factors. We have analyzed our 7-year data set on methane emissions from rice fields in Tu Zu, China, to delineate the relationships between emissions and a number of variables that were measured at the same time. Our work was done in fields that were managed under prevailing agricultural practices of the region. Consequently, only the effect of factors that vary from year to year or during the growing season can be calculated. In our study we measured the effects of environmental variables (soil temperature, wind speed, sky cover) and agricultural factors (planting density, water level, rice cultivars, organic fertilizer amounts, yield). Of these variables, soil temperature had the most significant effect on methane emissions resulting in Q10 values of about 2 (1.5--3). The effect of sky cover, and even water levels, was to change the soil temperature, which in turn affected the methane flux. Wind tended to increase emissions, possibly by agitation of the soil. Of the agricultural variables, planting density had the most significant but complex effect on methane emissions. We studied emissions from up to 4 times the normal planting density under otherwise similar agricultural conditions in the same fields. For a four fold increase in planting density the seasonal average emissions increased by about a factor of 2. Rice cultivars had a small but detectable effect. The amount of organic fertilizer and the yields did not affect methane emissions in our fields. The lack of an effect from the fertilizers is attributed to a saturation phenomenon whereby methane emissions do not respond to continual increases in organic material after some sufficiently high level. ¿ 1998 American Geophysical Union