We used whole-tree, open-top chambers to expose 13-year-old loblolly pine (Pinus taeda L.) trees, growing in soil with high or low nutrient availability, to either ambient or elevated (ambient + 200 mumol mol(-1)) carbon dioxide concentration ([CO2]) for 28 months. Branch growth and morphology, foliar chemistry and gas exchange characteristics were measured periodically in the upper, middle and lower crown during the 2 years of exposure. Fertilization and elevated [CO2] increased branch leaf area by 38 and 13%, respectively, and the combined effects were additive. Fertilization and elevated [CO2] differentially altered needle lengths, number of fascicles and flush length such that flush density (leaf area/flush length) increased with improved nutrition but decreased in response to elevated [CO2]. These results suggest that changes in nitrogen availability and atmospheric [CO2] may alter canopy structure, resulting in greater foliage retention and deeper crowns in loblolly pine forests. Fertilization increased foliar nitrogen concentration (N-M), but had no consistent effect on foliar leaf mass (W-A) or light-saturated net photosynthesis (As,,). However, the correlation between A(sat) and leaf nitrogen per unit area (N-A = WANM) ranged from strong to weak depending on the time of year, possibly reflecting seasonal shifts in the form and pools of leaf nitrogen. Elevated [CO2] had no effect on W-A, N-M or N-A, but increased A,,, on average by 82%. Elevated [CO2] also increased photosynthetic quantum efficiency and lowered the light compensation point, but had no effect on the photosynthetic response to intercellular [CO2], hence there was no acclimation to elevated [CO2]. Daily photosynthetic photon flux density at the upper, middle and lower canopy position was 60, 54 and 33%, respectively, of full sun incident to the top of the canopy. Despite the relatively high light penetration, WA, NA, A(sat) and R-d decreased with crown depth. Although growth enhancement in response to elevated [CO2] was dependent on fertilization, [CO2] by fertilization interactions and treatment by canopy position interactions generally had little effect on the physiological parameters measured.