The reasons why only a small proportion of African children infected with Plasmodium falciparum develop severe or fatal malaria are not known. One possible reason is that children who develop severe disease have had less previous exposure to malaria infection, and hence have less acquired immunity, than children who develop a mild clinical attack. To investigate this possibility we have measured titres of a wide range of anti-P. falciparum antibodies in plasma samples obtained from children with severe malaria, children with mild malaria and from children with other illnesses. Mean antibody levels in patients with malaria were higher than those in patients with other conditions but, with only one exception, there were no significant differences in antibody titres between cases of severe or mild malaria. A parasitized-erythrocyte agglutination assay was used to estimate the diversity of parasite isolates to which children had been exposed; plasma samples obtained from children with cerebral malaria recognized as many isolates as did samples obtained from children with mild disease. Our findings do not provide any support for the view that the development of severe malaria in a small proportion of African children infected with P. falciparum is due to lack of previous exposure to the infection.