A second generation assay for antibody to hepatitis C virus (anti-HCV) was used to screen 78 southern Indian individuals with a high risk of infection. RT-PCR targeted at the 5' end untranslated region (5'UTR) of the HCV genome was used to evaluate evidence of viraemia in 32 anti-HCV positive sera. The PCR products amplified from the 5'UTR of the HCV genome from 24 patients were sequenced, revealing the existence of two distinct groups of sequences: 21 corresponded to HCV type 1 while the other three sequences had 95% to 99% identity to HCV type 3. Two of these three isolates had more than 90% nucleotide identity in the NS5 region to established 3b sequences whereas the other had less than 74% nucleotide identity to any of the published genotype 3 (3a, 3b, 3c, 3d, 3e and 3f) sequences. However, a search of the EMBL nucleotide database revealed 91% identity to the unpublished sequence of an isolate of HCV from Indonesia. We provide evidence that these two isolates may represent a novel subtype within genotype 3. Our data also suggest that HCV genotype 1 predominates over HCV genotype 3 in southern India.