Dengue has been known for more than 200 years. The first dengue viruses were isolated about 50 years ago. Prior to the 1950's, dengue was considered a mild febrile disease, though rare hemorrhagic and fatal cases were known to occur. After that date, the first epidemics of dengue hemorrhagic fever (DHF) appeared in Southeast Asia, and DHF became the most important cause of childhood morbidity and mortality in the region. The emergence of DHF epidemics was first explained by mutations affecting dengue viruses, making them more virulent, but this hypothesis was not retained. Then, the "secondary infection" or "immune enhancement" theory was proposed to explain the increased virulence of dengue viruses when children had a secondary infection. This second hypothesis is still actually favoured. However, observations in Southeast Asia, some Pacific islands, and Americas do not agree with the "secondary infection" hypothesis, which consequently has been modified several times. Recent advances in molecular biology have led to the recognition that some viral strains are more virulent than others. Another hypothesis is the selection of more virulent dengue strains by the new vector Ae. aegypti, replacing the local vector Ae. albopictus, when urbanization and modern transportation increased in Southeast Asia after the last war. Comparisons between epidemics are very difficult, because of the distinction between DHF cases according to WHO criteria and dengue fever (DF) cases with hemorrhages. This distinction has no pathogenic or prognostic grounds, and makes the task of clinicians more difficult. The actual situation in countries facing dengue epidemics makes clear that this disease will continue to be a public health problem for some time to come.