In a controlled trial of weekly malaria chemoprophylaxis with chloroquine and pyrimethamine there were no significant differences in type and frequency of severe morbidity during chemoprophylaxis. Administration of chemoprophylaxis during the current and immediately preceding month was associated with significantly fewer episodes of severe morbidity in the chloroquine and pyrimethamine groups when each was compared with the control multivite group. After chemoprophylaxis had been stopped, significantly more episodes of severe morbidity occurred in the chloroquine group than the control group, but a similar trend in the pyrimethamine group was not statistically significant. In the control group most of the episodes of severe morbidity, including those episodes which were associated with heavy parasitaemia, occurred below the age of 4 years. In contrast, the children who received chemoprophylaxis continued to experience such illness at older ages. The difference between the chloroquine group and the control group in respect of age at time of severe morbidity was statistically significant.