Five cultivation implements (two flex-tine cultivators, brush hoe, and rolling and shovel cultivators) were evaluated for weed control and effect on broccoli, snap bean, and sweet corn yield. Each implement was used alone or in combination with flex-tine cultivators. Standard broadcast herbicide treatments were: broccoli, metolachlor at 2.3 kg ai/ha posttransplant; snap bean, metolachlor preemergence (PRE) at 2.3 kg ai/ha followed by fomesafen postemergence (POST) at 0.4 kg ai/ha; sweet corn, atrazine at 1.1 kg ai/ha plus metolachlor at 2.3 kg ai/ha PRE. In the snap bean trials, an additional treatment consisted of metolachlor at 2.3 kg ai/ha PRE applied in a 25-cm band over the crop row followed by fomesafen POST at 0.4 kg ai/ha directed at the crop row, with a subsequent shovel cultivation. In the sweet corn trials, all cultivation treatments received an application of atrazine at 1.1 kg ai/ha plus metolachlor at 2.3 kg ai/ha PRE in a 25-cm band over the crop row. In broccoli, weed control and yields with all cultivation treatments were comparable to those with broadcast herbicide. In snap bean, flex-tine cultivators failed to control weeds and prevent yield reductions compared to the broadcast herbicide treatment. When either the brush hoe or the shovel cultivator was preceded by flex-tine cultivation, weed control and snap bean yield were comparable to those with broadcast herbicides. Banded herbicides followed by shovel cultivation also controlled weeds and prevented yield reductions. In sweet corn, banded herbicides controlled in-row weeds, and despite variable between-row weed control, all cultivation treatments yielded as well as the broadcast herbicide treatment. Relative performance of the cultivators was influenced by soil type and rainfall as well as crop canopy characteristics.