Although the hemagglutination-inhibition (HI) test has been the standard test used by the World Health Organization (WHO) for the classification of serologic response in dengue infections, it is slow, requiring paired specimens. Furthermore, not all investigators have accepted the classification. Thus, there is an urgent need to develop a rapid test which employs a single specimen. We developed an enzyme-linked immunosorbent assay (ELISA) for rapid classification of serologic responses in dengue infections based on the ratio of IgM and IgG in a single specimen. Using the criteria established by the WHO (1986) for comparison, concordant results were obtained in 81% and 95% of primary and secondary infections, respectively, when serum specimens were tested as pairs. When tested as single specimens, the diagnoses by ELISA and HI agreed in 41% and 52% of acute specimens of primary and secondary infections, respectively. The lower rate of concordance in acute-phase samples was due to the absence of detectable IgM in acute specimens collected at outpatient clinics. On the other hand, diagnoses by ELISA and HI agreed in 79% and 95% of primary and secondary infections when single convalescent specimens were used. Analysis of the discordant results between the two tests revealed that the interpretation by the IgM-IgG ratio agreed better with HI classifications practised by some investigators than it did with the WHO.