The objective of this study was to determine the relative importance of different factors in controlling the amount and partitioning of gaseous nitrogen losses from denitrification in an agricultural soil. Denitrification (nitrous oxide + dinitrogen) and nitrous oxide emissions were measured on repacked soil cores using acetylene inhibition. Both the amount and partitioning of gaseous nitrogen losses could best be explained by consideration of the relative supply of, and demand for, terminal electron acceptors (TEAs). Increasing soil water content and carbon addition, which limit oxygen supply, increased denitrification through increased demand for alternate TEAs. Nitrate addition had no effect on denitrification except when soil nitrate concentration was low. The primary product of denitrification was nitrous oxide except where high carbon availability (i.e. high demand for TEAs) was present in combination with low soil nitrate concentration (i.e. low alternate TEA supply), resulting in increased reduction of nitrous oxide.