Patterns of evolution of N2O and N-2 due to denitrification in intact cores of a clay loam soil were measured using a He/O-2 atmosphere ''flow-over'' incubation system housed in a temperature-controlled room. Square section cores were taken from a grassland site in SW England under extensive grazing management and assembled into composite turves, each comprising 25 cores in a 5 x 5 array, which were placed in each of six incubation vessels. After replacement of N-2 in the soil pores with He, the headspace gas above each turf was continuously flushed with a stream of 20% O-2 in He, which was directed to either waste or dual gas chromatographs. The effects of the major controls on denitrification were investigated while simulating the application of NO3- fertilizer to the sward made via a N-2-free irrigation assembly placed above each incubation vessel. Denitrification increased with increasing NO3 added within the range equivalent to 0-150 kg ha(-1), and with increasing water-filled pore space within the range 70-90%. The denitrification response to variation in the other controls did not agree well with the results of previous studies: although the initial rate of denitrification increased with a Q(10) of 2 within the range 5-30 degrees C, there was no clear trend in the total N denitrified at temperatures above 10 degrees C; denitrification decreased with increasing soil pH within the range 5.1-9.4. The N2O-to-N-2 ratio increased with increasing NO3-, and with decreasing water content, pH and temperature. Antecedent soil aerobicity also had a large effect on the N2O-to-N-2 ratio: after 7 d of either aerobic or anaerobic conditioning, the ratio was 1.74 or 0.15, respectively. In most of the experimental runs, less than 100%, and sometimes less than 50%, of the added N could be accounted for in gaseous products. The results indicate the need to develop and apply techniques that enable concurrent measurement of all relevant processes of N transformation, such as assimilatory NO3- reduction, nitrification and plant uptake, if prediction of denitrification in field soils is to be improved. (C) 1997 Elsevier Science Ltd.