The objective of this work was to evaluate the performance of nonprotected (by fences or pickets) seedlings of leguminous tree species and an equitable mix of these species, introduced in pastures of Brachiaria brizantha cv. Marandu and Panicum maximum cv. Tanzânia, in the presence of cattle. The experimental design was a completely randomized one, in a 2x5 factorial arrangement, with two grasses (marandu and tanzânia) and four leguminous species (Mimosa artemisiana, Pseudomanea guachapele, Enterolobium contortisiliquum, Acacia farnesiana), and a mix of these species, with three replicates. Evaluations were made for: seedling height, stem diameter, canopy diameter, seedling survival, pasture frequency and ant occurrence. The statistical difference between the means of the main canonical variable, by the test of Scott-Knott, indicated the formation of three groups, standing out the one formed by M. artemisiana and mixture of leguminous in the two pastures, plus E. contortisiliquum and A. farnesiana in pastures of marandu and tanzânia grass, respectively. Differences between the means of treatments for each variable, calculated by means of Bonferroni's intervals of confidence, showed that M. artemisiana presented the greatest height and seedling survival in the pasture of marandu grass. This leguminous seedlings, without protection, are indicated to be introduced in pastures of marandu grass in the region, in the presence of cattle.
Brachiaria brizantha; Mimosa artemisiana; Panicum maximum; grazing; silvipastural system; survival
