This study investigated the structure of, and carbon dioxide fluxes at, a 'primary' rain forest (PRF) in SW Amazonia, Brazil, and a distributed secondary rain forest (SRF) in SW Cameroon. The total above-ground biomass and leaf area index (LAI) at PRF were respectively 220 (±95% c.l. 48) Mg ha^-1 and 4.0 (±95% c.l. 0.7) m² m^-2, and at SRF 90 (±95% c.l. 9.4) Mg ha^-1 and 4.4 (±95% c.l. 0.9) m² m^-2. A novel method was devised to quantify the vertical profile in LAI: SRF was distinguished from PRF by a higher concentration of leaf area near the ground. Three methods were used to determine the flux of CO₂ from soil and gave overall agreement (static and dynamic chambers, and eddy covariance). The mean soil efflux in PRF and SRF was respectively 5.5 μmol m^-2 s^-1 (±95% c.l. 0.2; n = 42) and 4.5 μmol m^-2 s^-1 (±95% c.l. 0.2; n = 178) at 20 - 24 °C. The temperature response was higher in PRF than SRF (Q₁₀ = 2.3 vs 1.9). Soil efflux rates were also obtained from cerrado vegetation in central Brazil, where the efflux was 3.2 μmol m^-2 s^-1 (±95% c.l. 0.2; n = 10) the Q₁₀ 1.6, at 16 - 23 °C. Heterogeneity in emission was higher in SRF than in PRF and could be described by a non-linear model incorporating the variables: soil temperature, organic carbon and total nitrogen (r² = 0.82). Carbon was the most important variable determining respiration in SRF; soil moisture was not limiting. There was no observable effect of season on efflux rates in either rain forest, but a decline occurred in cerrado during the dry season. Effluxes of CO₂ were measured from stems and branches of diameter 0.002 m - 1.6 m in 24 species in PRF and 17 species in SRF; emission rates were 0.1 - 3.3 μmol m^-2 s ^-1 with a Q₁₀ of 1.8 in PRF, and 0.2 - 5.2 μmol m^-2 s ^-1 with a Q₁₀ of 1.6 in SRF. Maintenance respiration was 80% and construction respiration 20% of total woody tissue respiration (R_t) in SRF. A functional model described the relationship between R_t and diameter in SRF better than a purely empirical one (r² = 0.66). A novel method was devised in estimate sap CO₂ concentrations which in SRF were 1.2 - 11.0 nmol 1^-1 for Distemonanthus benthamianus and Musanga cecropioides. Sap CO₂ levels were highly sensitive to sap pH, and represented 1 - 30% of cuvette-measured leaf photosynthesis. Maintenance leaf respiration (R_m) was measured through the vertical profile during the night in PRF and SRF. R_m increased with height at both sites: ˜0.2 in PRF vs ˜0.3 in SRF at 1.5 m and ˜0.5 in PRF vs ˜0.9 in SRF at 26 m (values normalised to 22 °C, units: μmol m^-2 s^-1). Leaf nitrogen and potassium concentrations (N_leaf and P_leaf) declined with height in the canopy. P_leaf concentrations were higher in SRF than PRF where P_leaf appeared to limit respiration. R_m was not significantly related.