A simple top-down model of canopy photosynthesis (P) was developed and tested in this study. The model (referred to as the Q(e)-MM model) is P = alphaQ (e) P (max)/(alphaQ ( e ) + P (max)), alpha and P (max) are quantum-use efficiency and potential P, respectively. Q (e) is given by Q (d) (0) + kQ (b) (0), where Q (d) (0) and Q (b) (0) are the diffuse and direct photosynthetically active radiation (PAR) incident on the canopy, respectively. Q (e) can be considered to be the effective incident PAR contributing to P and k is a measure of the contribution of Q (b) (0) to Q (e). When k = 1, the Q(e)-MM model becomes the regular Michaelis-Menten type model of P (referred to as the MM model). A major objective of this study was to determine how well the Q(e)-MM model could estimate P of a 56-year-old coastal Douglas-fir stand. To this end, we parameterized the Q(e)-MM model using five and half years of eddy-covariance measurements of CO(2) flux above the Douglas-fir stand. The Q(e)-MM model, with the incorporation of a function of air temperature, accounted for 74% of the variance in over 34,000 half-hourly P measurements. P estimated using the Q(e)-MM model had no systematic errors with respect to Q (d) (0). Although the Q(e)-MM model has only one more parameter than the MM model, it accounted for 30% more variance in P than the latter when total incident PAR exceeded 900 micromol m(-2) s(-1). On average, k was found to be 0.22. We show that this small value of k reflects the significant effect of the scattering of the solar beam and the fraction of light-limited sunlit leaves. We also show that the success of the Q(e)-MM model was due to the fact that a large fraction of the sunlit leaves were light-limited as a result of their orientation to the solar beam.