Nitrous oxide (N2O) and carbon dioxide (CO2) fluxes were measured in a boreal forest during two growing seasons with soil gradient and chamber methods. N2O fluxes obtained by these two techniques varied from small emission to small uptake. N2O fluxes were of the same order of magnitude, however, the fluxes measured by the soil gradient method were higher and more variable than the fluxes measured with chambers. The highest soil gradient N2O fluxes were measured in the late summer and the lowest in the autumn and spring. In the autumn, litter fall induced a peak in N2O concentration in the organic O-horizon, whereas in the spring N2O was consumed in the O-horizon. Overall, the uppermost soil layer was responsible for most of the N2O production and consumption. Soil gradient and chamber methods agreed well with CO2 fluxes. Due to the very small N2O fluxes and the sensitivity of the flux to small concentration difference between the soil and the ambient air, the flux calculations from the O-horizon to the atmosphere were considered unreliable. N2O fluxes calculated between the soil A- and O-horizons agreed relatively well with the chamber measurements.