New Zealand’s agriculture is heavily dependent upon the symbiotic fixation of atmospheric nitrogen. Consequently little fertilizer nitrogen is used in pastoral farming. Nevertheless, N deficiency is widespread throughout New Zealand pastures, indicating that either the N supply is inadequate, or that significant losses are occurring from the system. The contribution of denitrification to these losses is unknown, primarily because at present there is no suitable technique available for directly measuring denitrification losses in the field. The major gaseous losses from the soil due to denitrification occur as nitrogen and nitrous oxide. Detection of nitrous oxide evolution is possible by several methods including gas chromatography, mass spectrometry and infra-red analysis. Several techniques have been proposed for direct N₂0 measurements in the field. Measurement of the relatively small quantity of nitrogen evolved from soils in the presence of the atmospheric background has proved difficult for many years and is a significant barrier to the determination of N losses. The objectives of this thesis were to: 1) develop a laboratory method to identify soils from which large denitrification losses may be expected, 2) develop a technique for direct measurement of denitrification losses in the field, and 3) gain some understanding of the significance and occurrence of the denitrification activity occurring in New Zealand soils. A laboratory incubation technique was developed to measure the relative denitrification potential of soils. The denitrification potential is defined as the maximum rate of nitrate dissimilation under anaerobic conditions without addition of exogenous reductant. The technique was based on the acetylene inhibition of nitrous oxide reductase. Soils were amended with 250 μg N0₃-N.g⁻¹ and incubated at 25°C and 100% water holding capacity for eight hours in an argon atmosphere containing 10% acetylene. The incubation atmosphere was sampled hourly for N₂0 production. Six soils in duplicate can be assayed by one person in the course of a normal working day. A widespread (approximately ten-fold) variation in denitrification potential was found between soils. Soils exhibiting the lowest denitrification potentials were also seen to exhibit low nitrification activities indicating that denitrification was limited by the availability of nitrate substrate. As the availability of nitrate substrate increased (increased nitrification activity), other factors influenced the denitrification potential and there was no relationship between denitrification and nitrification activity. Soils containing allophane generally exhibited high denitrification activities. The denitrification potential also varied with stock rates, higher denitrification potentials being observed under high stocking rates on a given soil. This effect was attributed to the increased availability of C and N substrate through excreta return by the animals. Denitrification potentials were also influenced by the type of crop. Soils under maize crops exhibited markedly lower denitrification potentials than adjacent soils in pasture. This effect was attributed to the depletion of available carbon by the maize crops. The maximum denitrification potential in a soil always occurred in the 0-30 mm layer. This association of the denitrification potential with the rhizosphere could not wholly be explained on the basis of nutrient availability. Below 30 mm the denitrification potential decreased markedly, even in the presence of adequate C and N substrate. Addition of further C and N substrate at depth did not stimulate increased denitrification activity. Addition of cow’s urine to a soil was not accompanied by an increase in the denitrification potential of that soil even though the C and N substrate levels increased many-fold. The soil was demonstrated to already be at its maximum denitrification potential in situ at the time of the experiment. After three weeks, a covered plot of the same soil exhibited a marked decrease (40%) in the denitrification potential. It appears that this decrease resulted from some enzymic inhibition or toxicity related to the addition of urine. Storage techniques reported for soils used in denitrification studies vary widely. Soils are stored air dry at room temperature, at 4°C and frozen, for periods of a few days to severa1 months. It was demonstrated in this work that the denitrification potential of a soil is significantly affected by even a few days’ storage at 4°C. This effect was attributed to a rapid decline in available carbon. Some soils demonstrated two linear phases of denitrification during an eight hour incubation. It was shown that the first phase was unaffected by chloramphenicol (a protein synthesis inhibitor) while the transition to the second phase did not occur in the presence of chloramphenicol. The first phase was attributed to the activity of pre-existing enzymes (i.e. in situ) while the second phase was taken as an indication of the maximum denitrification potential of that soil due to stimulation of increased enzyme activity. The maximum denitrification potential of a soil is relatively stable seasonally, decreasing markedly only during the peak summer months. This effect was attributed to the effect of low soil moisture on microbial processes. While the soil was denitrifying at the maximum potential the denitrifying enzymes were in a state of full induction in situ. Groundwater nitrate concentrations are high in shallow aquifers in the Waikato. This implies that denitrification in this region is limited more by available carbon than nitrate supply. To facilitate the direct measurement of nitrogen evolution under field conditions a lysimeter was specially designed and constructed. The purpose of the lysimeter was to achieve and maintain a low nitrogen background within a soil core. By flushing the soil core with nitrogen-free gas, nitrogen concentrations between 2000 and 5000 ppm v/v could easily be achieved. A unique feature of this lysimeter was the ability to form an artificial barrier at the base of the soil core using a flow of nitrogen-free gas. The lysimeter walls were double skinned, permitting a flow of gas to exit through holes in the inner wall near the base. This barrier proved effective in preventing the inward diffusion of soil air during measurements of nitrogen evolution. By introducing into the soil core a low concentration of nitrogen enriched in N-15, nitrogen evolution could be monitored using a high precision double beam Micromass 602c mass spectrometer. By measuring the isotopic dilution of the N-15 spike, δ¹⁵N changes representing rates of nitrogen evolution as low as 10 kg N/ha/yr could be measured. Nitrous oxide evolution can be measured at the same time by switching the Micromass to the single beam mode. Sample pre-treatment is kept to a minimum. The lysimeter-¹⁵N technique allows the first direct single sample measurement of nitrogen and nitrous oxide evolution in the field. The lysimeter was also shown to be effective in achieving and maintaining high acetylene concentrations in a soil core. This capability meant that field estimates of denitrification using only the evolution of nitrous oxide were also possible. The lysimeter offers an advantage over previous field acetylene techniques in that it does not suffer from the uneven acetylene distribution associated with the multi-point acetylene source technique. The lysimeter also allows the oxygen concentration within the soil to be closely controlled. The contribution of Rhizobia to the leguminous fixation of atmospheric N is well known. It was demonstrated that free-living Rhizobia japonicum are capable of denitrification in soils under laboratory conditions. Even under moderate oxygen tensions (10% v/v), Rhizobia appear to remain as active denitrifiers. The question of the significance of Rhizobia denitrification in the field is raised in the light of previous results which indicated a marked preference for denitrification to occur in the rhizosphere, and the notable reduction of denitrification potential in soils of non-legumous crops.