From June 1, 1994 to May 31, 1995 a total of 24,700 cases of dengue (7.01/1,000 population) were reported to the laboratory-based surveillance system in Puerto Rico (1991-1994, annual average: 2.55/1,000). Dengue virus 2 predominated. The earliest indicator of epidemic activity was the virus isolation rate in May 1994 (14.0% versus 5.7% average). The male-to-female ratio among cases was 1:1.1; 65.4% were younger than 30 years (the 10 to 19 year age group had the highest incidence, 11.8/1,000). At least 5,687 cases (23.0%) showed a hemorrhagic manifestation; 4,662 (18.9%) were hospitalized, and 40 died (0.2%; 10 laboratory-positive). Two cases documented by laboratory were transmitted by unusual routes--intrapartum and through a bone marrow transplant. Among 2,004 hospitalized cases reported by infection control nurses, 139 (6.9%) fulfilled the criteria for dengue hemorrhagic fever (DHF) and another 13 cases (0.6%) had dengue shock syndrome. This epidemic produced the largest number of hospitalizations, DHF cases, and deaths from any dengue epidemic in Puerto Rico. Severity did not change throughout the year. Surveillance capabilities were maintained by temporary, simplified reporting methods, none of which could be recommended as the single method of choice for surveillance; each must be used (on site, or as a service available from a reference laboratory) at the right time in the epidemic cycle. The utility of comparisons of current and previous data underscores the value of long-term surveillance. Our analysis was unable to document whether significantly increased transmission occurred more often in cities where the water supply was rationed or where the local landfill was closed.