Using a ground-based and an above-canopy eddy covariance system in addition to stem respiration measurements, the annual respiratory fluxes attributable to soil, stems and foliage were determined for a Scots pine (Pinus sylvestris L.) forest growing in central Siberia. Night-time foliar respiration was estimated on the basis of the difference between fluxes measured below and above the canopy and the stem respiration measurements. Comparison of the effects of night-time turbulence on measured CO 2 fluxes showed flux loss above the canopy at low wind speeds, but no such effect was observed for the ground-based eddy system. This suggests that problems with flow homogeneity or flux divergence (both of which would be expected to be greater above the canopy than below) were responsible for above-canopy losses under these conditions. After correcting for this, a strong seasonality in foliar respiration was observed. This was not solely attributable to temperature variations, with intrinsic foliar respiratory capacities being much greater in spring and autumn. The opposite pattern was observed for stem respiration, with the intrinsic respiratory capacity being lower from autumn through early spring. Maximum respiratory activity was observed in early summer. This was not simply associated with a response to higher temperatures but seemed closely linked with cambial activity and the development of new xylem elements. Soil respiration rates exhibited an apparent high sensitivity to temperature, with seasonal data implying a Q 10 of about 7. We interpret this as reflecting covarying changes in soil microbial activity and soil temperatures throughout the snow-free season. Averaged over the two study years (1999 and 2000), the annual respiratory flux was estimated at 38.3 mol C m −2 a −1 . Of this 0.61 was attributable to soil respiration, with stem respiration accounting for 0.21 and foliar respiration 0.18.