Human monkeypox is a zoonosis occurring sporadically in the tropical rain forest of western and central Africa. The exact incidence and geographical distribution are unknown, since many cases are not recognized. Special surveillance was established in three regions in Zaire in 1981 that led to a substantial increase in reported cases. The question arose as to the possibility that clinical diagnostic errors cause some cases of monkeypox to be misdiagnosed as other eruptive diseases. This paper presents the results of a study assessing the extent of and reasons for these clinical diagnostic errors in areas where health staff as well as the general public are aware of human monkeypox. In Zaire in the period 1981-1986, 977 persons with skin eruption not clinically diagnosed as human monkeypox were laboratory tested. 3.3% of human monkeypox cases were found among 730 patients diagnosed as cases of chickenpox, 7.3% among cases diagnosed as "atypical chickenpox" and 6.1% among cases with skin rash for which clinical diagnosis could not be established. The diagnostic difficulties were mainly based on clinical features characteristic of chickenpox: regional pleomorphism (in 46% of misdiagnosed cases), indefinite body-distribution of skin eruptions (49%), and centripetal distribution of skin lesions (17%). Lymph-node enlargement was observed in 76% of misdiagnosed patients. In the absence of smallpox, the main clinical diagnostic problem is the differentiation of human monkeypox from chickenpox. The presence of lymphadenopathy, pre-eruptive fever and slower maturation of skin lesions are the most important clinical signs supporting correct diagnosis of monkeypox.