Introducing nitrogen (N)-fixing legumes into cereal-based crop rotations reduces synthetic fertiliser-N use and may mitigate soil emissions of nitrous oxide (N 2O). Current IPCC calculations assume 100% of legume biomass N as the anthropogenic N input and use 1% of this as an emission factor (EF)-the percentage of input N emitted as N 2O. However, legumes also utilise soil inorganic N, so legume-fixed N is typically less than 100% of legume biomass N. In two field experiments, we measured soil N 2O emissions from a black Vertosol in sub-tropical Australia for 12 months after sowing of chickpea ( Cicer arietinum L.), canola ( Brassica napus L.), faba bean ( Vicia faba L.), and field pea ( Pisum sativum L.). Cumulative N 2O emissions from N-fertilised canola (624 g N 2O-N ha -1) greatly exceeded those from chickpea (127 g N 2O-N ha -1) in Experiment 1. Similarly, N 2O emitted from canola (385 g N 2O-N ha -1) in Experiment 2 was significantly greater than chickpea (166 g N 2O-N ha -1), faba bean (166 g N 2O-N ha -1) or field pea (135 g N 2O-N ha -1). Highest losses from canola were recorded during the growing season, whereas 75% of the annual N 2O losses from the legumes occurred post-harvest. Legume N 2-fixation provided 37-43% (chickpea), 54% (field pea) and 64% (faba bean) of total plant biomass N. Using only fixed-N inputs, we calculated EFs for chickpea (0.13-0.31%), field pea (0.18%) and faba bean (0.04%) that were significantly less than N-fertilised canola (0.48-0.78%) ( P<0.05), suggesting legume-fixed N is a less emissive form of N input to the soil than fertiliser N. Inputs of legume-fixed N should be more accurately quantified to properly gauge the potential for legumes to mitigate soil N 2O emissions. EF's from legume crops need to be revised and should include a factor for the proportion of the legume's N derived from the atmosphere.
