Wood productivity in plantation forests depends on successful choices of genetic materials and silvicultural practices. Wisely defining the intitial tree spacing according to edaphoclimatic conditions is important as spacing influences the amount of water, nutrients and light available for each tree. Resource availability, in turn, influences some important processes controlling wood production such as tree survival, CO2 uptake, and partitioning of photosynthates to stem wood and other tree compartments. Different genotypes may respond differently to tree spacing, due to possible differences in their canopy structure, physiology, and strategies for soil ressources acquisitions and carbon allocations. Our objective in this study was to investigate the interactive effects of genotype and tree spacing on wood production of fast-growing clonal eucalypt plantations in Southeastern Brazil. Tree growth, gross primary production (GPP) and partitioning of GPP to total belowground carbon flux (TBCF), wood and leaf production, and aboveground respiration were assessed in an experimental trial involving three commercial E. urophylla clones (AEC 0144, AEC 224 and COP 1404) planted at two spacings (3 m x 2 m; 3 m x 4 m). We hypotetized that (1) the most productive genotypes are those with both the highest GPP and the highest carbon partitioning to stem wood production; and (2) at closer spacing, leaf area index (LAI) and GPP are higher, but the partitioning to wood production is lower, and the belowground partitioning is higher, due to reduced soil water and nutrients availability per tree. Biomass and LAI were measured at ages 40 months (mid-rotation) and 52 months, and GPP, TBCF, wood production and leaf production were estimated between these two stand ages. Stem biomass and LAI were higher, for all clones, at the closer spacing (3 m x 2 m), as we could have expected since this spacing (corresponding to a stocking of 1667 tree ha- 1) is the most commonly chosen in industrial eucalypt plantations in this region. Consistently with clonal ranking established in other studies, at spacing 3 m x 2 m clones were ranked as AEC 0144 > AEC 224 > COP 1404 for their wood production, but this ranking almost vanished at wide spacing. Our first hypothesis was supported by data: at spacing 3 m x 2 m the most productive clones (AEC 0144 > AEC 224 > COP 1404) were those with both the highest GPP (6.14 > 4.63 > 4.57 kg C m- 2 year- 1) and the highest carbon partitioning to wood production (0.33 > 0.29 > 0.26), although the differences between COP 1404 and AEC 224 were not significant (at P = 0.05). Our second hypothesis was only partially supported by data: wood biomass at age 40 months was higher at the closer spacing, but this positive effect of tight spacing seemed to decrease with stand age, as wood production and GPP measured between ages 40 and 52 months remained higher at the closer spacing only for clone AEC 0144. There was no consistent difference in the partitioning of GPP to wood production between the two spacings. The agerelated decrease of the positive effect of high stockings may result from a faster depletion of deep soil water reserves in dense stands.