Naturally acquired transmission-blocking immunity to Plasmodium vivax was studied in three groups of patients from the southern coast of Mexico: primary cases (Group A, 61% of the study population), secondary cases with the prior infection seven or more months earlier (Group B, 23%), and secondary cases with the previous malaria experience within six months of the present study (Group C, 16%). Anopheles albimanus mosquitoes were fed with patients' infected blood cells in the presence of autologous or control serum, with or without heat-inactivation. Patients from all three groups had transmission-blocking immunity, although the quality and quantity of this blocking activity was significantly higher in the two secondary patient groups (B and C). Only primary malaria cases produced transmission-enhancing activity (23% of the cases), which was dependent on heat-labile serum components. The levels of patient group transmission-blocking immunity and mosquito infectivity were used to calculate the probabilities of a mosquito becoming infective after taking a blood meal from a P. vivax-infected patient from any one of the three groups. This probability was 0.025, with Group A patients providing the major source of these infections (92% risk from Group A and 4% risk for Groups B and C).