Subsistence farmers with little access to costly farm inputs such as pesticides and fertilizers must rely on natural pest regulation and cultural controls. To assess the effects of different management practices on a key pest, bean fly densities were compared among 19 farmers' field in central Malawi. The close proximity of these fields, synchrony of planting dates, and similar field sizes allowed a rigorous, empirical comparison of the relative susceptibility of different bean varieties, the effects of soil fertility, and effects of surrounding vegetation. The two most commonly used varieties differed significantly in bean fly infestation levels. Total soil nitrogen was positively correlated with the most common species of bean fly (Ophiomyia spencerella); and soil phosphorus levels were negatively correlated with population densities of both bean fly species. Similar rates of biological control of bean flies by parasitic wasps occurred in all fields. Complementary experiments with potted plants showed a direct relationship between synthetic NPK fertilizer level and O. spencerella density. However, the effect was reversed in treatments with soluble fish protein fertilizer. In addition, straw mulching of newly sown fields, a practice not used by farmers in the study, may reduce bean fly damage.