The results of a 16-year, parasitological and clinical, follow-up study on bancroftian filariasis in three villages in north-eastern Tanzania are reported and analysed with respect to the long-term development of microfilaraemia and disease. The study revealed an extraordinary static situation, both at the community level and at the individual level. No significant differences were observed in any of the three villages, with respect to total or age-specific prevalences of microfilaraemia, hydrocele in males aged > or = 20 years or leg elephantiasis in all subjects aged > or = 20 years, when the results from 1975 were compared with those from 1991. Overall, 542 of the subjects (37.2% of the total) examined in 1975 were re-identified and re-examined in 1991. Of those microfilaraeic in 1975, 81.9% were also microfilaraemic in 1991. Similarly, 81.3% of those amicrofilaraemic in 1975 were also amicrofilaraemic in 1991. Thus, most subjects had the same microfilarial (mf) status at both surveys. The low mf loss rate indicates that re-infection of microfilaraemic individuals commonly takes place, and results in repeated sequences of microfilaraemia in the same individuals. The rate of gain of microfilaraemia was independent of age, thus indicating no age-related change in the examined population's susceptibility to develop or sustain microfilaraemia. Individual differences in susceptibility to develop mirofilaraemia therefore appear to be innately determined, or to be acquired in pre-natal or early post-natal life. Neither the presence of hydrocele in 1991, nor the development of new cases of hydrocele over the 16-year period were related to the mf status in 1975 or 1991, and no association between microfilaraemia and this chronic clinical manifestation was observed. The association between leg elephantiasis and mf status could not be analysed because of the small sample size.