BACKGROUND Hemagglutination (HA) and hemagglutination inhibition (HI) assays are conventionally used for detection and identification of influenza viruses. HI assay is also used for detection of antibodies against influenza viruses. Primarily turkey or chicken erythrocytes [red blood cells (RBCs)] are used in these assays, as they are large, nucleated, and sediment fast, which makes it easy to determine the titer. Human influenza viruses agglutinate RBCs from chicken, human, and guinea pig, but not from horse. Human influenza viruses bind preferentially to sialic acid (SA) linked to galactose (Gal) by α 2, 6 linkage (SA α 2, 6-Gal), whereas avian influenza (AI) viruses bind preferentially to SA α 2, 3-Gal linkages. With this background, the present study was undertaken to study erythrocyte binding preferences and receptor specificities of AI viruses isolated from India. METHODS A total of nine AI virus isolates (four subtypes) from India and three reference AI strains (three subtypes) were tested in HA and HI assays against mammalian and avian erythrocytes. The erythrocytes from turkey, chicken, goose, guinea pig and horse were used in the study. The receptor specificity determination assays were performed using goose and turkey RBCs. The amino acids present at 190 helix, 130 and 220 loops of the receptor-binding domain of the hemagglutinin protein were analyzed to correlate amino acid changes with the receptor specificity. RESULTS All tested highly pathogenic avian influenza (HPAI) H5N1 viruses reacted with all five types of RBCs in the HA assay; AI H9N2 and H5N2 viruses did not react with horse RBCs. For H5N1 viruses guinea pig and goose RBCs were best for both HA and HI assays. For H9N2 viruses, guinea pig, fowl and turkey RBCs were suitable. For other tested AI subtypes, avian and guinea pig RBCs were better. Eight isolates of H5N1, one H4N6 and one H7N1 virus showed preference to avian sialic acid receptors. Importantly, two isolates of HPAI H5N1, H9N2 and H11N1 viruses showed receptor specificity preference to both avian and mammalian sialic acid (α-2, 3 and α-2, 6) receptors. CONCLUSIONS Use of different types of RBCs resulted in titer variations in HA and HI assays. This showed that RBCs giving optimum HA and HI titers would increase sensitivity of detection and would be more appropriate for identification and antigenic analysis of AI viruses. Analysis of 16 amino acids in the receptor-binding domain of the hemagglutinin of HPAI H5N1 viruses revealed that the only variation observed was in S221P amino acid position. Two H5N1 viruses showed S221P amino acid change, out of which only one H5N1 virus showed preference to α 2, 6 sialic acid receptor. One H5N1 virus isolate with amino acid S at 221 position, showed preference to α 2,3 as well as α 2,6 sialic acid receptors. This indicated that factor(s) other than S221P mutation in the hemagglutinin are probably involved in determining receptor specificity of H5N1 viruses. This is the first report of receptor specificity and erythrocyte binding preferences of AI viruses from India.