A study is described in which private physicians participated in the surveillance and diagnosis for dengue disease within a community. Using simple, economical methods it was shown that a significant proportion of specimens submitted were positive for dengue infection. Of 610 fingerprick blood specimens collected on filter paper strips, 7% to 33% were positive by serology. Of these specimens, 423 (69%) were single specimens and 92 (22%) of the single specimens were positive by serology. Of 153 fingerprick blood specimens collected in heparinized, glass capillary tubes, 15% to 50% yielded a dengue virus isolate. When a positive result was based on serology and/or virus isolation, an overall positive rate of 39% was obtained for one of the participating clinics. Inclusion of thrombocytopenia and leukopenia in selection of cases significantly improved the positive rate. It is argued that this simple and practical community-based approach will result in more accurate surveillance of dengue transmission within an endemic area and thus aid in the development of long-term strategies for control of disease.