In order to clarify the mechanism of increased methane emission from rice paddy fields under the continuation of no-till direct seeding cultivation, we measured methane flux from rice paddy fields with gray lowland paddy soils in Okayama Prefectural General Agricultural Center, by using a closed chamber method, where no-till direct seeding cultivation was continued for several years. In addition, we measured the dissolved methane concentration in the soil solution of the plowed layer and in the flooded water, the total amount of dissolved methane existing in the plowed layer, and the biomass of rice plants above the ground to understand the relationship between the biomass and methane emission. Consequently, the following results were obtained. 1) Methane emissions in the no-till direct seeding rice paddy field (ND), which was continued over eight years, became comparable to or higher than those in the tilled and transplanting rice paddy field (TT). And methane flux became much higher in ND than in TT, in the period from the flooding onset to the maximum tillering stage. The plant biomass in that period was usually larger in ND than in TT, because direct seeding in ND occurred about one month before transplanting in TT. In the later period, however, methane flux in ND was almost equal to or a little bit lower than that in TT. 2) Methane concentration in the upper layer (0-5cm) of the plowed soil layer was much higher in ND than in TT, whereas that in the lower layer (5-15cm) of the plowed layer was much higher in TT than in ND. The total amount of methane in all the plowed layer was much higher in TT than in ND, in the early and later stage of the rice growing season. 3) In contrast, dissolved methane concentration in standing water was much higher in ND than in TT through out the rice growing season. 4) The thickness of organic matter layer accumulated on the soil surface in ND increased year by year, reaching 20mm after a ten-year continuation of ND. 5) In the early stage of the rice growing season, the roots in ND were concentrated in the upper layer (0-5cm) of the plowed soil layer, compared to those in TT, although in the later stage, the growing roots in ND were also sufficiently extended through the plowed layer. 6) These results indicate that, in the early stage of rice cultivation, methane production in the plowed soil layer in ND occurred much more in the upper layer where organic matter was accumulated and the soil became more reductive than in TT. Thus, methane formed in the upper plowed layer in ND is transported efficiently to the atmosphere, because the roots were more extended in the upper layer of the plowed layer, and the biomass of rice plant was also larger in ND than that in TT. 7) In addition, the methane production rate in the upper layer in ND increased year by year with the increase in the organic matter layer where the root biomass also increased, resulting in increase in the methane emission rate in ND with the continuation of no-till direct seeding cultivation. 8) The methane emission rate in the rice paddy field converted to TT from ND after eight continuous years, did not increase for the first rice growing season, compared to that in ND which was not converted to TT.