In highly malaria-endemic areas, thick smears are usually positive regardless of the clinical context. Therefore the simple positivity or negativity of the thick smear is not an adequate criterion for distinguishing malaria from other causes of fever. In order to define simple parasitological and clinical criteria for diagnosing clinical malaria with a small risk of error, a study was undertaken in a rural area in the Congo where malaria transmission is intense and perennial. Results of the systematic determination of the parasite density of 1562 samples from persons of all ages considered representative of the population of the studied area are compared to those from 327 febrile patients, 204 patients detected during medical consultations held in the villages and 123 febrile schoolchildren detected during surveys for fever. The analysis of the clinical data and the parasitological results clearly demonstrates the importance of the parasite density determination for the diagnosis of clinical malaria. Clinical Plasmodium falciparum malaria is unlikely to occur in children under 15 years if the parasite/leucocyte ratio is less than 1.5. On the contrary this diagnosis is very probable if the parasite/leucocyte ratio is higher than 2. Clinical criteria were too non-specific to serve as useful diagnostic criteria.