N turnover in flooded rice soils is characterized by a tight coupling between nitrification and denitrification. Nitrification is restricted to the millimetre-thin oxic surface layer while denitrification occurs in the adjacent anoxic soil. However, in planted rice soil O(2) released from the rice roots may also support nitrification within the otherwise anoxic bulk soil. To locate root-associated nitrification and denitrification we constructed a new multi-channel microelectrode that measures NH(4)(+), NO(2)(-), and NO(3)(-) at the same point. Unfertilized, unplanted rice microcosms developed an oxic-anoxic interface with nitrification taking place above and denitrification below ca. 1 mm depth. In unfertilized microcosms with rice plants, NH(4)(+), NO(2)(-) and NO(3)(-) could not be detected in the rhizosphere. Assimilation by the rice roots reduced the available N to a level where nitrification and denitrification virtually could not occur. However, a few hours after injecting (NH(4))(2)HPO(4) or urea, a high nitrification activity could be detected in the surface layer of planted microcosms and in a depth of 20-30 mm in the rooted soil. O(2) concentrations of up to 150 mu M were measured at the same depth, indicating O(2) release from the rice roots. Nitrification occurred at a distance of 0-2 mm from the surface around individual roots, and denitrification occurred at a distance of 1.5-5.0 mm. Addition of urea to the floodwater of planted rice microcosms stimulated nitrification. Transpiration of the rice plants caused percolation of water resulting in a mass flow of NH(4)(+) towards the roots, thus supporting nitrification.