An epidemiological study was conducted in Eastern and Northern India to determine the genomic diversity of rotaviruses in these parts of the country. In 2001, a total of 126 Group A rotavirus positive samples were detected from children below 4 years of age with diarrhoea from Kolkata, Dibrugarh and Bhubaneswar in Eastern India, and Chandigarh, a city in Northern India. All the samples were genotyped for VP7 (G-type) and VP4 (P-type) gene by reverse transcription (RT) and multiplex PCR using different type specific primers. The strains with G1P[8] (32.5%) was predominant as reported earlier [Das et al. (2002) J Clin Microbiol 40:146-149] followed by G2P4 and only one sample was of G4P[8] specificity. Along with these common types some rare strains like G1P[6], G2P[8], G2P[6], G4P[4], and G4P[6] were also detected in 14.3% of cases. Thirty percent of samples in this study were mixed infections and 21 (16.7%) specimens remained untypeable either for the VP7 or for the VP4 gene. After sequencing of the VP7 gene, two G9 strains (RMC321 and ISO-3) were identified with P[8] and P[19] specificities. Sequence analysis revealed that they have much lower homology to the G9 strains (116E, INL1, and G16) isolated earlier from Indian subcontinent, but have much higher homology to isolates from Argentina, Brazil, Malawi, Taiwan, and USA suggesting a separate progenitor for these strains.