Several cellular and humoral mechanisms probably play a role in natural immunity to Plasmodium falciparum malaria, but the development of an effective vaccine has been impeded by uncertainty as to which antigens are targeted by protective immune responses. Experimental models of malaria have shown that cytotoxic T lymphocytes (CTL) which kill parasite-infected hepatocytes can provide complete protective immunity against certain species of Plasmodium in mice, and studies in The Gambia have provided indirect evidence that CTL play a protective role against P falciparum in humans. By using an HLA-based approach, termed reverse immunogenetics, we have previously identified peptide epitopes for CTL in liver-stage antigen-1 and the circumsporozoite protein of P falciparum. We have extended this work to identify CTL epitopes for HLA class I antigens that are found in most individuals from Caucasian and African populations. Most of these epitopes are in conserved regions of P falciparum. CTL peptide epitopes were found in a further two antigens, thrombospondin-related anonymous protein and sporozoite threonine and asparagine rich protein, indicating that a subunit vaccine designed to induce a protective CTL response may need to include parts of several parasite antigens. However, CTL levels in both children with malaria and in semi-immune adults from an endemic area were low suggesting that boosting these low levels by immunisation might provide substantial or even complete protection against infection and disease.