Recently, Tsukiyama-Kohara et al. (1993) reported that most hepatitis C viruses (HCVs) in Japan can be classified into two types, type 1 and type 2, on the basis of the NS4 region nucleotide sequence. They developed a new assay in which antibodies against group-specific recombinant proteins of the NS4 region were measured by ELISA (serologically defined genotype, serotype). In the present study, we examined 306 patients with chronic liver disease due to HCV infection. The sensitivity of this assay was 98.7% (302). The serotype distribution of HCV was230 (75.2%) for type 1, 65 (21.2%) for type 2, 7 (2.3%) for mixed, both being positive, and 4 (1.3%) indeterminate. The frequency of type 1 was significantly higher than that of type 2 (P < 0.01). There were no significant differences in clinical characteristics among the mixed and the indeterminate serotypes. Among the mixed-serotype patients, 4 (57.1%) showed seroconversion to a single serotype at 6 and 9 months later, although the serotypes of the indeterminate-serotype patients were also indeterminate at 6 and 9 months later.Using aliquots of the same serum samples, HCV genotyping was carried out by the reverse transcription polymerase chain reaction (RT-PCR) method using type-specific primers derived from the NS5 region of HCV to verify the specificity of this serotyping. The sensitivity of genotyping by RT-PCR was 167 (91.3%). The HCV genotypes determined by both methods were consistent in 161 (88.0%) of the cases, and there were no contradictory results for any sample between the two methods. These findings indicate that serological genotyping might be useful in determining the HCV genotype among Japanese patients with HCV infection.