For decades, large areas of native forests have been and still systematically destroyed in Brazil, many of them leading to forest plantations stands for industrial purposes. The adequacy to the recent environmental legislation regarding the maintenance of permanent preservation areas and legal reserves required reversing the areas occupied with commercial plantations into native forests. This study aimed at verifying the effect of techniques to accelerate the natural regeneration of native vegetation in abandoned production stands of Eucalyptus saligna. The study was conducted in areas belonging to the Suzano Papel e Celulose, at Parque das Neblinas, Bertioga, SP. The experiment was set up in a commercial Eucalyptus saligna stand with 45 ha, third rotation, previously managed in seven-year cutting cycles. The experimental design was in randomized blocks with six treatments and four replications. The treatments were: T1= control; T2-T5= reduction of 25, 50, 75 and 100% of basal area, respectively, through girdling and poisoning and T6= clearcut. Each plot had 20 x 40 meters. An inventory on the Eucalyptus saligna population was performed, as well as a phytosociological survey of the entire natural regeneration over 1.3 m in height 1,417 individuals of Eucalyptus saligna were measured and showed an absolute density of 738.02 ind/ha and 22.68 m2/ha basal area. As to the treatment applied (T2-T5) to the eucalyptus individuals, the method was proved efficient. We sampled 5,863 individuals of native vegetation belonging to 135 species of 77 genera and 39 families. The total density estimated was 1,035.41 individuals/ha and the basal area was 6.39 m2/ha for the DBH ≥ 5 cm class and 3,864.58 individuals/ha and basal area of 3.25 m2/ ha for the natural regeneration with DBH < 5 cm. There was no significant effect of the treatments at a 5% probability level for the variables studied. Although, the richness analysis corrected (correction through the Fisher’s α index) for the DBH ≥ 5 cm class showed decrease in the control plots between the initial and final phases, and an increase in the richness of the remaining treatments. In addition, the percent variation of the Shannon’s diversity index between the initial and final times had a significant effect for treatments and blocks, being the T3 showed significant increase in species diversity (H’). Although no significant effect was observed for the basal area, a negative linear correlation of this variable with the eucalypt basal area was observed. The result points to a trend of increase in the basal area of the native vegetation with DBH > 5 cm as the eucalypt basal area is reduced. T3 (50 % reduction of the eucalypt basal area by poisoning) showed best results for density, basal area and species richness of for both natural regeneration classes after 40 months.
