Clinical immunity to malaria was studied by quantifying the intensity of symptoms as well as by measurement of several hematologic indicators of pathology (the erythrocyte sedimentation rate [ESR], serum bilirubin, reticulocyte count, plasma tumor necrosis factor-alpha [TNF-alpha], and blood glucose levels) in 39 Plasmodium vivax malaria patients exposed to endemic malaria in southern Sri Lanka, and for comparison in 43 nonimmune patients who were residents of nonmalarious regions of the country. The intensity of 11 symptoms was scored numerically in all patients using a questionnaire. This clinical score was validated by introducing internal controls to the questionnaire, and by correlating it with the underlying pathology. Both the intensity of clinical disease as well as the degree of underlying pathology were found to be significantly lower in endemic area patients (mean clinical score = 8.8, median ESR = 8 mm) compared with the nonendemic area patients (mean clinical score = 19.0, median ESR 31.5 mm). Endemic area patients also had lower parasite densities (mean = 0.06%) than those from the nonendemic area (0.12%) (P < 0.05). However, at any parasite density, both clinical disease and pathology were significantly less in the endemic area patients (P < 0.001, for both clinical score and ESR), indicating that the clinical immunity seen in the endemic area patients was a true tolerance of parasites. Although plasma TNF-alpha levels were elevated in both groups of patients, they were significantly higher in the nonendemic area patients than in patients from the endemic area (P < 0.01). Furthermore, at comparable levels of plasma TNF-alpha, nonendemic area patients had both a higher intensity of clinical disease and an underlying pathology than those from the endemic area, suggesting that if TNF-alpha is indeed a mediator of clinical disease, the endemic area patients may be tolerant to its effects. Hypoglycemia was not observed in any of these P. vivax patients despite some with high levels of plasma TNF-alpha.