The general objective of the present work was to quantify the volumetric, biomass and carbon stoks in different compartments of a natural forest, in the Atlantic Forest bioma. The work was developed in a forest fragment with 17 hectares, named Mata da Silvicultura, located in the municipality of Viçosa-MG, Brazil, pertaining to the Universidade Federal de Viçosa. In the first chapter the characterization and the floristic and phytosociologic analyses of plants were done. In the second chapter the equations for the volume, biomass and carbon of the stem of sampled trees were estimated. In the third chapter the volume, biomass and carbon stoks of the stems of the individuals with DBH = 5 were estimated and of the biomass and carbon stocks of the understory vegetation, of the litter and roots were estimated. Three approach levels were used, with 15 plots in each level, in the sampled forest inventory. In Level I, 20 x 50 (0.1 ha) plots were used and all individuals with DBH = 5 were measured; in Level II all individuals with DBH < 5 cm and stem height = 1.3 m, were weighed which were found inside a sub plot of 5 m x 5 m (25 m2); and, in Level III all material laid on the ground inside the sub plots of 2.5 m x 2.5 m (6.25 m2) was weighed. From the inventory data the floristic and phytosociologic analyses were done and the results were used in the selection of the species to be scaled. Samples of the wood and bark were collected for the estimation of the biomass and carbon content. Three non linear regression models were evaluated to estimate the volume, biomass ad carbon of the stem, outside and inside bark. The best equations were used to produce the estimations of the stocks of wood, biomass and carbon. After the analyses it was concluded that: the Mata da Silvicultura has a high diversity of species; from the 10 species with the greatest importance value-IVI, almost all of them also presented the greatest volumetric importance value-VIV; the use of IVI or VIV provides a list of species with different positions; the arboreous individuals have stems with an average height of 7.1 m and total average height of 10.18 m; the amount of carbon, both in the wood and in the bark had smaller estimation than 50% of the biomass value; the equations related to the Schumacher and Hall model, with the use of the independent variables DBH and stem height were the ones that best adjusted to the observed data of volume, biomass and carbon of the stems of the trees, outside and inside bark; the total mean volume for DBH = 5 cm was estimated in 281.51 m3 ha-1, and 15.2% of this value correspond to the bark; the total mean biomass was estimated in 227.40 t ha-1, and, in relation to the live trees (DBH = 5 cm) the stock was estimated in 188.16 t ha-1 (82.8%), the dead trees (DBH = 5 cm) in 8.01 t ha-1 (3.5%), the non arboreous species (DBH = 5 cm) in 9.57 t ha-1 (4.2%), the small trees (DBH < 5 cm and Hf = 1.3 m) in 6.67 t ha-1 (2.9%), the seedlings (Hf < 1.3 m) in 3.37 t ha-1 (1.5%) and the litter in 11.62 t ha-1 (5.1%); for the mean total stock of carbon estimation is of 108.98 t ha-1, and the estimations for the live trees (DBH = 5 cm) of 90,1 t ha-1 (82.6%), the dead trees (DBH = 5 cm) of 3.81 t ha-1 (3.5%), the non arboreous species (DBH = 5 cm) of 4.64 t ha-1 (4.2%), the small trees (DBH < 5 cm and Hf = 1.3 m) of 3.24 t ha-1 (3.0%), the seedlings (Hf < 1.3 m ) of 1.64 t ha-1 (1.5%) and the litter of 5.64 t ha-1 (5.2%). The methods traditionally used in forest mensuration to estimate the carbon stock in commercial eucalypt plantations and other species were efficient to estimate the volumetric, biomass and carbon stocks in different compartments of the forest.
