Pre-emergence activity of pendimethalin on propanil-resistant jungle rice (Echinochloa colona) was demonstrated in glasshouse trials. Both susceptible and resistant populations, collected from Costa Rica, were controlled by 1É25 kg ha~1, the usual application rate used in the Ðeld where Rottboellia cochinchinensis is also a problem. When applied post-emergence, propanil performance was improved by the addition of low doses of pendimethalin to the herbicide mixture. A propanil-resistant selection was controlled by 0É23 kg ha~1 pendimethalin ] 0É54 kg ha~1 propanil at the one-to-two leaf stage, and 0É23 kg ha~1 pendimethalin ] 1É08 kg propanil at the three-to-four leaf stage compared to 1É08 kg and 2É16 kg ha~1 respectively when propanil was applied alone. This suggests that pendimethalin improves post-emergence control in the Ðeld compared to the standard propanil treatment and can provide residual preemergence control of late-germinating individuals, so reducing the propanil selection pressure. For e †ective jungle rice control, growers apply propanil (3É84 kg ha~1) at 10 and 20 days after planting (DAP) followed by one application of fenoxaprop-P-ethyl (0É045 kg ha~1) at 35 DAP. Field experiments, conducted in dry-seeded upland rice in southern Costa Rica, demonstrated that under high jungle-rice population pressure, one application of pendimethalin at 1É5 kg ha~1 provided an e †ective replacement for propanil, resulting in reduced weed-control costs.