Denitrification was investigated during 1985 and 1986 in a stagnogleyic brown earth (Macmerry) and a stagnogley (Winton), both soil types widely used for cereal production in south-east Scotland. The crop was winter barley. Much of the work was devoted to the evaluation and attempted field application of the acetylene-inhibition technique. This involves blocking the terminal reduction of nitrous oxide to elemental nitrogen by acetylene concentrations in excess of 0.1% by volume; total denitrification loss is then estimated by measuring the resultant enhanced nitrous oxide flux at the soil surface. For prolonged periods during the growing season relatively high concentrations of nitrous oxide in the soil atmosphere indicated significant rates of denitrification. It proved impossible, however, to obtain an adequate distribution of acetylene throughout the profile, resulting in improbably low denitrification estimates. This finding casts doubt on any reported measurements using the acetylene- inhibition technique in wet soils with poor structure. The method was applied with more success on a freely-drained Darvel soil. An alternative technique was investigated for the gleyed soils: laboratory incubation of intact cores in an atmosphere containing acetylene. This produced increased estimates of denitrification but was found to introduce other uncertainties. It was concluded that further development was required before the technique could be used to give reliable quantitative results. Measured gaseous diffusion constants and nitrous oxide concentrations at depth were employed in the Fick?s Law calculation of nitrous oxide flux; calculated losses during 1986 ranged from 0.2-0.5 kg N /ha for direct-drilled Winton plots to 3.6-7.4 kg N /ha for normally-ploughed M acmerry soil. These were minimum values for total gaseous loss, since losses as elemental nitrogen were not recorded. A mathematical model of denitrification was developed. It depends on numerical solution of the differential equations governing the simultaneous steady-state diffusion and reduction of oxygen, nitrate, and nitrous oxide in a spherical m icro-environm ent (an aggregate where aggregates are present - an "effective aggregate" in a structureless soil). A model aggregated soil is pictured as an assembly of spherical aggregates with log-normally distributed aggregated radius and reductive potential; the radius of the effective aggregate in a structureless soil is determined by the density of air-filled pores. Intra?aggregate diffusion constants are calculated by a method which amounts to an assumption of parallel diffusion through all possible serial combinations of intra-aggregate pores. Reductive potentials for the various reactions considered in the model are assumed to be proportional to one another. The model predicts an approximately linear relationship between the denitrification rate of an assembly of soil aggregates and its anaerobic fraction calculated according to the method of Smith (1980); such a relationship has in fact been observed (Parkin and Tiedje, 1984). Model results illustrate the importance of soil structure: a model clay soil continues to denitrify at moisture potentials much lower (more negative) than a model sand. Calculated whole-soil denitrification rates range from 0 g N/ha/d to 5.2 kg N/ha/d. Reported field measurements range from 0 g N/ha/d to 4.5 kg N/ha/d.

The linkage between denitrification activity, N gas emissions, and the size of the denitrifier community in soils of an upland pasture was investigated. Special emphasis was placed on soil pH as a regulating factor, the spatial distribution of denitrification, and the degree of cattle impact. The thesis has been based on field and laboratory measurements using both conventional and modern methods of soil ecology.