Two field experiments were conducted to determine the soil persistence, crop toxicity, and effectiveness of four herbicides in controlling purple nutsedge (Cyperus rotundus L.). Excellent control of purple nutsedge was obtained with 2,6-dichlorobenzonitrile (dichlobenil) and good control was obtained with N-hydroxy-methyl-2,6-dichlorothiobenzamide. Both 3-iert-butyl-5-chloro-6-methyluracil (terbacil) and 5-bromo-3-teri-butyl-6-methyluracil were effective in controlling purple nutsedge, but maximum control was not obtained until 9 to 12 months after these chemicals were incorporated in the soil. Toxicity of all chemicals to beans (Phaseolus vulgaris, var. humilis L., Blue Lake), corn (Zea mays var. rugosa L., Golden Security), summer squash (Cucurbita moschata Duchesne, var. Yellow Crookneck), tomato (Lycopersicum esculentum Mill., var. Flora-del), and gladiolus (Gladiolus hortulanus Bailey, var. Friendship) increased as the rates were increased and decreased with time after treatment. Infestation of purple nutsedge from escaping tubers increased as the herbicidal persistence decreased.