Fluctuations of soil temperature may greatly affect the estimate of N mineralization and nitrification based on mean temperatures, particularly when the daily temperature amplitude is large as in many tropical soils. I investigated the effects of fluctuating temperature on N mineralization and nitrification in an Oxisol, and established a method that takes this effect into account by using current information obtained from experiments carried out at constant temperatures. Soil incubations were made with and without NH4 addition at three constant temperatures (20, 30 and 40 degreesC) and three fluctuating temperature patterns (FTP). The mean temperature of the FTP was 30 degreesC. Two FTP had a large temperature amplitude (8 h at 20 degreesC, 8 h at 30 degreesC, 8 h at 40 degreesC, FTP1, and 12 h at 20 degreesC, 12 h at 40 degreesC, FTP2) and the third was an actual pattern observed at the field with a near-sinusoidal wave and a small amplitude (27.1-33.2 degreesC, FTP3). Nitrogen mineralization obeyed first order kinetics and nitrification was linear against time with an initial period of NH4 immobilization. Nitrogen mineralization was higher for FTP 1 and FTP2 than for FTP3 or constant 30 degreesC. The temperature response of the constant rate of N mineralization was fitted to the Q 10 equation. Calculations performed with the parameters obtained at constant temperature applied to the FTP described the experimental data well. This suggests that the differences between FTP 1, FTP2 and constant 30 degreesC were due to the non-linear response of the constant rate of N mineralization to temperature. In other words, microbial activity at a given temperature was the same under FTP and at constant temperature. Similar results were obtained for nitrification but the differences between FTP and constant 30 degreesC were less. Daily mean temperature was not suitable for estimating daily N mineralization and nitrification when the temperature amplitude was large and I propose a method based on the use of daily mean rates. For a small daily temperature amplitude, the use of the daily mean temperature seems appropriate because it results in only a small bias which is less than the uncertainties associated with the estimate of the rates of soil microbial processes. (C) 2002 Elsevier Science Ltd. All rights reserved.