Cotton yield response was evaluated in 1987, 1988, and 1989 when weeds were managed with preemergence fluometuron [none (0%), band (50%), or broadcast (100%) surface coverage], cultivation (none, one, two, or three times), and postdirected fluometuron + MSMA. Weed densities (primarily prickly sida, morningglories, and hemp sesbania) varied widely among years and were directly related to early season rainfall. Postdirected herbicide application or cultivation(s) had little effect on weed density. The use of a banded fluometuron application reduced weed biomass 28 to 47%. A further decrease was observed when preemergence fluometuron was increased from banded to broadcast coverage. Seed cotton yields were low with no preemergence fluometuron. Banded fluometuron and at least one cultivation had yields similar to broadcast fluometuron only. Cotton yields were related to weed density and weed biomass in a hyperbolic relationship. Low weed densities caused more yield loss per unit weed density than higher densities.
