Soil-active herbicides were investigated for control of live oak (Quercus virginiana Mill.) on the Texas Coastal Prairie and Claypan areas with and without mechanical treatment. Roller chopping before, during, or after herbicide application had very little influence on mortality of live oak compared to the herbicide applied alone. Roller chopping did allow increased grazing of livestock for at least 3 yr after knockdown of the brush. Tebuthiuron {N-[5-(1,1-dimethylethyl)-1,3,4-thiadiazol-2-yl]-N,N′-dimethylurea} was generally the most effective herbicide for live oak control; bromacil (5-bromo-3-sec-butyl-6-methyluracil), karbutilate [tert-butylcarbamic acid ester with 3(m-hydroxyphenyl)-1,1-dimethylurea], and picloram (4-amino-3,5,6-trichloropicolinic acid) were somewhat less effective, especailly at 2.2 kg/ha. Limited data suggested that hexazinone [3-cyclohexyl-6-(dimethylamino)-1-methyl]-1,3,5-triazine-2,4(1H,3H)-dione] also was highly effective for control of live oak. Primary invaders on either mechanically or chemically treated areas were American beautyberry (Callicarpa americana L.), southern dewberry (Rubus trivialis Michx.), boneset (Eupatorium perfoliatum L.), yankeeweed (Eupatorium compositifolium Walt.), brownseed paspalum (Paspalum plicatulum Michx.), and carpetgrass (Axonopus affinis Chase). A herbicide was generally equally effective for live oak control whether applied broadcast or in rows as granules, or applied as sprays to the soil surface or subsurface.