Tuberculosis (TB) is still a major cause of morbidity and mortality. It is clear that control requires more than simple availability of antibiotics. In order to gain insight into the disease, DNA fingerprinting has been applied to the study of bacterial population structure. This technology has been used to quantitate various components of the disease in a high-incidence community, viz. recent transmission (RT) and reactivation (RA) and to monitor these over time as a tool to quantitate changes in the epidemic. In our high-incidence community, we find unexpectedly high strain diversity, lower than predicted RT, and that reactivation disease dominates. This technology can be used to examine and challenge traditional dogmas. Quantitative measure of RT varies over time, using a two-year sliding window for estimation as a useful period. The results show that the "epidemic" consists of subepidemics characterized by strain families that wax and wane in the community of TB patients. The technology is shown to be a useful and quantitative tool to assess disease status and can therefore be used to monitor intervention strategies and refine and monitor results of new control measures.