We aimed to evaluate the selectivity of the herbicides imazapyr, glyphosate, metribuzin, and sulfentrazone, when applied on four native species of semideciduous forests from southeastern Brazil: Acacia polyphylla and Enterolobium contortisiliquum (Fabaceae), and Ceiba speciosa and Luehea divaricata (Malvaceae). We did a randonmized design with four replicates per treatment: the application of imazapyr (125, 250, 500, 1000, 2000 and 3000 g ha-1), sulfentrazone (100, 200, 400, 800, 1400 and 2400 g ha-1), glyphosate (90, 180, 360, 720, 1440 and 2160 g ha-1), and metribuzin (360, 720, 1,920, 2880, 5760 and 8400 g ha-1), besides the control (without any herbicide). We evaluated phytotoxicity symptoms 30 days after application (DAA), and also dry biomass of leaves. Experimental plot comprised one seedling with 30 cm height. We concluded that glyphosate showed the lowest selectivity in relation to four tree species; therefore, we suggest that its application could be done through direct aplications. Behavior of four tree species varied in relation to selectivity of herbicides. For C. speciosa, imazapyr was the most selective herbicide, followed by sulfentrazone. Metribuzin was selective for A. polyphylla. In relation to E. contortisiliquum, metribuzin presented the lowest phytotoxicity, followed by sulfentrazone. For L. divaricata, only sulfentrazone showed selectiveness.
competition; degraded area; phytotoxicity; selectiveness; herbicides; tropical forest
