Molecular genetic analyses of P. falciparum have led to the cloning and sequencing of a number of antigens that are potential candidates for vaccination against malaria. Seroepidemiological studies in endemic areas have attempted to assess the relative importance of these antigens in protection against malaria. In this paper, we attempt to evaluate the relative contributions of conserved and strain-specific immune responses by modelling their influence of age-specific patterns of infection and disease. The modelling exercises in this paper clearly demonstrate that the observed patterns of age-prevalence are best explained by proposing that the accumulation to a threshold of an immune response against a conserved determinant is required for protection against infection, while 'anti-disease' immunity develops more linearly with exposure. This is compatible with the conjecture that the parasite population is structured into several independently transmitted strains, that each confers some degree of 'anti-disease' immunity, but does not protect against further infection by the same strain. Within this framework, the average duration of parasitaemia increases with age, as previously encountered strains endure for longer periods at a subclinical level. Indirect evidence for the increase in duration of parasitaemia with age may be obtained from a comparison of age-prevalence curves between dry and rainy seasons. By using mathematical methods to structure epidemiological and immunological information, we provide a coherent theoretical framework for the dissection of the important components of naturally acquired immunity to malaria.