Twenty herbicide treatments were evaluated on conventional-till (plow, disc, and harrow), minimum-till (disc only), and no-till planted soybeans [Glycine max (L.) Merr.] from 1976 through 1978 at the Belleville Research Center in St. Clair County, Illinois. The soil type was a Weir silt loam (Typic Ochraqualf) characterized by poor internal drainage and 1.2% organic matter. Weed population by species, weed control, and soybean population, injury, and yield were obtained. Fall panicum (Panicum dichotomiflorum Michx.) and giant foxtail (Setaria faberi Herrm.) were the dominant species in all tillage systems, exceeding 1 million plants/ha in the conventional and no-till plots. These species and ivyleaf morningglory [Ipomoea hederacea (L.) Jacq.] were the most difficult to control each year. Weed control was the poorest in the no-till plots because of the large size of the weeds at the time of herbicide application, insufficient rainfall following, and because the plots were not cultivated. The soybean population was equal in all tillage systems except in 1976 when the no-till population exceeded that in the other tillage systems. Treatments that included oryzalin (3,5-dinitro-N4,N4-dipropylsulfanilamide) caused 42 and 35% soybean injury in the 1976 minimum and no-till plots, respectively. Postemergence-applied naptalam (N-1-naphthylphthalamic acid) plus dinoseb (2-sec-butyl-4,6-dinitrophenol) caused leaf burn each year that ranged from 5 to 35% but all plants recovered within several weeks of application. The seedbed tillage method and herbicide treatments did not significantly affect soybean yields in 1976 when all the herbicides were effective. No-till yields in 1977 and 1978 were substantially lower than yields in conventional and minimum-till plots because of poor weed control. Soybean yields were 2506, 2466, and 1714 kg/ha in the conventional-till, minimum-till, and no-till plots, respectively, when averaged over the 3 yr and 20 herbicide treatments.