The efficacy of two strategies for control of Bancroftian filariasis using selective rather than community-wide diethylcarbamazine (DEC) chemotherapy was evaluated and compared in two endemic communities of north-eastern Tanzania, with pretreatment microfilariae (mf) prevalences of 22% and 38%, and geometric mean intensities (GMIs) of 668 mf/ml and 735 mf/ml of blood. All mf-positive cases in the first community were offered treatment with 6 mg of DEC/kg of body weight a day for 12 days (group 1), and those in the second community were offered treatment with two doses of 6 mg of DEC/kg of body weight at an interval of six months (group 2). The effect of treatment was followed both among those treated and at the community level. In treated individuals, there was a rapid decrease in the mf load that was significantly greater among those receiving the 12-day standard dose. One year after the start of treatment, the mf clearance rates were 59% and 39% and the GMIs were reduced by 99% and 97% among treated individuals in groups 1 and 2, respectively. However, at the community level, the mf prevalences were 16.3% and 27.9% (reduced by 27% and 26%) and the GMIs were 129 mf/ml and 224 mf/ml (reduced by 81% and 70%) one year after the start of treatment with the two regimens, respectively, suggesting that transmission continued at a significant level in the villages after treatment. The limitations of selective chemotherapy are discussed, and it is argued that strategies based on mass DEC chemotherapy would be more effective in reducing the microfilarial load in the community and thereby in reducing transmission.