Agricultural soils are a major source of N2O that has been linked to an enhanced greenhouse effect and the depletion of stratospheric ozone. The N2O produced in the soil is usually assumed to enter the atmosphere by diffusion. We postulated that significant amounts of N2O may also be emitted via plant transpiration, and conducted a controlled environment study to test this hypothesis. The shoots of canola (Brassica napus L.) were enclosed in chambers, and N2O emission from the foliage was quantified by measuring changes in N2O concentration following various soil treatments. At field capacity, a treatment not conducive to denitrification, there was no measurable emission of N2O from the plant foliage. At saturation, however, N2O concentration within the shoot chamber increased consistently with time. By the end of the sampling period, the N2O concentration in the shoot chamber was 0.45 L L-1, well above that in the soil head space (0.35 L L-1), demonstrating that N2O was released from the shoots and not the soil. Watering the soil with a N2O solution with a concentration of approximately 8 mL L-1 resulted in almost immediate release of N2O from the shoots, indicating that the N2O was not generated by the plants directly, but merely conveyed to the atmosphere by the transpiration stream. Experiments using barley (Hordeum vulgare L.) produced results similar to those with canola, suggesting that emission of N2O from plants may be a common phenomenon. These results demonstrate a mechanism for N2O emission from upland crops, and imply that measurement of N2O release solely from the soil surface may underestimate actual fluxes hom agroecosystems.