Of 4651 admissions between February 1995 and February 1996, 1043 had a presumed diagnosis of malaria. Six hundred and twenty-seven cases were confirmed by thick blood film examinations. The highest prevalence was in October (124/480 admissions) and the lowest in March (12/303). Sixty-five children died while 562 survived, 12 with defects. The first treatment in 422 children was chloroquine, in 143 quinine, in 59 halofantrin, and in three pyrimethamine with sulfadoxine (Fansidar). 23/422 patients started on chloroquine were switched to halofantrine, two to quinine. A higher mortality was associated with coma, convulsions, hepatosplenomegaly, pulmonary congestion, jaundice, haemoglobinuria, bladder paralysis, anuria. Anaemia and fever were more severe and hypoglycaemia more frequent in children who died than in children who survived (packed cell volume 18.5 +/- 7.1 per cent vs. 25.6 +/- 7.6 per cent, p < 0.001; temperature 39 +/- 1.1 degrees C vs. 38.7 +/- 0.9 degrees C, p < 0.05; random blood sugar < 40 mg/100 ml; 76 vs. 22 per cent, p < 0.01). There was no difference in the median age, pretreatment duration, and prevalence of diarrhoea and sickle cell disease. The male to female ratio was 1.5:1 in the surviving children vs. 1:1.03 in the dead.