Gas-exchange measurements on Eucalyptus grandis leaves and data extracted from the literature were used to test a semi-empirical model of stomatal conductance for CO2, g(sc)=g(0)+a(1)A/[(c(s)-Gamma) (1+D-s/D-0)], where A is the assimilation rate; D-s and c(s) are the humidity deficit and the CO2 concentration at the leaf surface, respectively; g(0) is the conductance as A-->0 when leaf irradiance -->0; and D-0 and a(1) are empirical coefficients. This model is a modified version of g(sc)=a(1)A h(s)/c(s) first proposed by Ball, Woodrow & Berry (1987, in Progress in Photosynthesis Research, Martinus Mijhoff, Publ., pp. 221-224), in which h(s) is relative humidity. Inclusion of the CO2 compensation point, Gamma, improved the behaviour of the model at low values of c(s), while a hyperbolic function of D-s for humidity response correctly accounted for the observed hyperbolic and linear variation of g(sc) and c(i)/c(s) as a function of D-s, where c(i) is the intercellular CO2 concentration. In contrast, use of relative humidity as the humidity variable led to predictions of a linear decrease in g(sc) and a hyperbolic variation in c(i)/c(s) as a function of D-s, contrary to data from E. grandis leaves. The revised model also successfully described the response of stomata to variations in A, D-s and c(s) for published responses of the leaves of several other species, Coupling of the revised stomatal model with a biochemical model for photosynthesis of C-3 plants synthesizes many of the observed responses of leaves to light, humidity deficit, leaf temperature and CO2 concentration. Best results are obtained for well-watered plants.