Field studies were conducted at Goiânia, GO, Brazil, on an Oxisol (clayey, kaolinitic, isothermic, Typic Haplustox) and at Jussara, GO, Brazil, on an Oxisol (loamy sand, kaolinitic, isothermic, Typic Haplustox) during 1995 and 1996 to determine the carryover effect of fomesafen, imazamox, and acifluorfen, applied to edible bean, on rotational crops (maize, sorghum, rice, and millet) and to estimate the level of soil residues under Brazilian Savanna conditions. Averaged across locale, year, and rate, fomesafen dissipation time (DT50) (37.5 d) was longer than acifluorfen (27.5 d) and imazamox (25.9 d). For both locations, soil herbicide persistence (average of herbicides) was longer in 1995 than in 1996. This was due to higher soil moisture content in 1996. The sensitivity of rotational crops to fomesafen and imazamox residues was, in decreasing order: sorghum, corn, millet, and rice, and for acifluorfen: sorghum, corn, rice, and millet. The period between herbicide application and rotational crop planting (PAP) varied in agreement with the sensitivity of rotational crops to herbicide residues in soil and the persistence of the herbicide. Considering both location and year, the PAP for fomesafen (250 g ai ha−1) ranged from 69 to 132 d for corn, 114 to 179 d for sorghum, 29 to 95 d for rice, and 52 to 111 d for millet; the PAP for imazamox (40 g ai ha−1) ranged from 68 to 111 d for corn, 78 to 139 d for sorghum, 25 to 75 d for rice, and 40 to 102 d for millet; and the PAP for acifluorfen (170 g ai ha−1) ranged from 56 to 89 d for corn, 96 to 139 d for sorghum, 61 to 95 d for rice, and 43 to 82 d for millet.