The trial was carried out during two sequential years in Ribeirão Preto Experimental Station, State of São Paulo, Brazil, in a Dusky Red Latosol cultivated with soybean IAC-11. Using a split-plot design; two dosages of linuron (1.00 and 2.00 kg/ha) and oryzalin (1.875 and 3.75 kg/ha) were compared with two control treatments: one inoculated with Bradyrhizobium japonicum strain SMS-443 (= 29W), and the other without inoculation. The experimental design was a randomized complete block-design with four replicates. The subplots were sampled on different days. The sensibility of Parker's and Santelman's methods for lab biossay was compared. The Parker's method performed more responsible for oryzalin while Santelman's for linuron. Under field conditions, the loss of activity of oryzalin was slow, persisting in the first year up to 56 and 84 days, for 1.875 and 3.75 kg/ha, respectively. In the second year, these dosages still decreased the plant-test growth in more than 50% at 84 and 145 days, respectively. Only after 390 days post applications oryzalin stopped affecting the plant growth. Although the initial phytotoxicity of linuron was very high, in the first year, there were no phytotoxicity symptom at 45 days in both dosages. In the second year, linuron was not detected at 28 days for the dosage of 1.0 kg/ha and at 70 days for the dosage of 2.0 kg/ha.
soybeans; herbicides; oryzalin; linuron; persistence