Observations are accumulating concerning: (1) the antigenic diversity of Plasmodium falciparum; (2) the diversity of the genetic immuno-competence of human hosts; (3) the associations between these diversities and disease, protection and defined immune responses; and (4) the possible mechanisms of acquired protection at different steps of the host-parasite interaction. These observations have led to speculations concerning the distribution of disease, the acquisition of protection, and vaccination prospects. Speculations on the latter have varied from pessimism about the possibility of developing a vaccine that would protect most people against most parasites, to optimism based on the hypothesis that a local P. falciparum population is composed of a few discrete subpopulations. If this hypothesis is correct, it may be possible to prevent most severe disease with a narrow-spectrum vaccination against a virulent minority of the subpopulations, or transmission could be interrupted by a relatively low coverage of a broad-spectrum vaccination. A conceptual model that might accommodate the observations is outlined. Its plausibility and testability are considered, as well as some of its implications for the planning and interpretation of epidemiological surveys and intervention trials, and perhaps for selection of antigens for inclusion in vaccines.