Soybean injury caused by the application of subdoses of 2,4-D or dicamba, in simulated drift.

2,4-D or dicamba can cause injuries and other deleterious effects on non-tolerant soybeans. Thus, the objective was to evaluate the potential for injury of subdoses of 2,4-D or dicamba, in drift simulation, for application in non-tolerant soybeans. Two experiments were carried out, one with 2,4-D and the other with dicamba. The treatments consisted of the application, in post-emergence of non-tolerant soybean, of subdoses 0; 1.35; 2.68; 5.37; 10.72; 21.45 and 42.9 g acid equivalent (ae) ha^-1 2,4-D choline salt or dicamba diglycolamine (DGA) salt. Injury symptoms in plants, plant height and yield were evaluated, and the results were subjected to regression analysis. Polynomial fit was possible for the doses of both herbicides, with deleterious effects on soybean, with reductions in height and yield. The application of 2,4-D ≥ 10.72 g ae ha^-1 was enough to cause injuries greater than 10% in plants, in simulated drift. The application of dicamba ≥1.35 g ae ha^-1 was enough to cause injuries greater than 30% in plants, in simulated drift. For both herbicides, greater potential for injury and reductions in soybean yield were observed for the application of the highest doses (21.45 and 42.9 g ae ha^-1).