We performed an open, randomized chemotherapy trial comparing the recommended first-, second- and third-line drug regimens, as well as mefloquine, for uncomplicated falciparum malaria in Bangladesh in 1996–1997. The regimens were chloroquine for 3 days (CQ, Group I), quinine sulphate for 3 days followed by single-dose sulfadoxine-pyrimethamine (Q3 + SP, Group II), quinine for 7 days (Q7, Group III), and mefloquine 20 mg/kg single dose (MEF, Group IV). Subjects were symptomatic patients, aged ≥12 years, with parasite density 500–250 000/mm³ and no history of taking antimalarials during the previous week. Drug administration was supervised and subjects were followed clinically and with blood slides in the hospital for 8 days, then as outpatients on days 14, 21 and 28. A total of 413 subjects (149, 145, 49 and 70 in Groups I–IV, respectively) completed the study. Early treatment failures (persistent or worsening clinical manifestations by day 3 confirmed with parasitological examinations) occurred only in the chloroquine group. RII and RIII parasitological failures occurred in 56%, 12%, 8% and 14% in Groups I–IV, respectively. There were significantly more clinical and parasitological failures with chloroquine than with Q3 + SP, which we now recommend as a better (but far from ideal) choice for first-line therapy. The alternative compounds show parasitogical evidence of Plasmodium falciparum resistance. Further studies are needed to determine the optimum treatment for malaria in Bangladesh.