We have attempted to provide a rational basis for improving the protocols for chemotherapy of malaria. We model the regression of parasitaemia by Plasmodium falciparum, its subsequent elimination from the body, or recrudescence, for populations of cells treated with chloroquine. Our model assumes that the drug forms a complex with some receptor in the parasite and that parasites possessing this complex die at a defined rate. We take into account that chloroquine is eliminated exponentially from the body. We show how the parameters of the model can be derived from observations in the field. The model correctly predicts the effects of drug dose, degree of initial parasitaemia, rate of parasite multiplication and degree of drug resistance to chloroquine chemotherapy. The level of parasitaemia will reduce to a minimum at sufficiently high concentrations of chloroquine, but only if the parasitaemia is reduced to below that of 1 parasite per infected person will a cure of malaria be obtained. Otherwise, recrudescence will, sooner or later, occur. We show that, even for drug-resistant malaria, if 2 doses of chloroquine are given to a patient with an interval of some 10 days between them, parasites can be eliminated from the body without toxic levels of chloroquine being reached.