The nucleotide sequences of the genes that code for the major inner capsid protein, VP6, of the human rotavirus strain 1076 (subgroup I), porcine rotavirus Gottfried (subgroup II), equine rotavirus strain H-2 (non-I/II) and equine rotavirus strain FI-14 (both subgroups I and II) have been determined. The sixth segment positive-stranded RNA encodes a protein of 397 amino acids in all strains with the exception of strain H-2 in which it encodes a protein of 399 amino acids. Alignment of amino acid sequences of the VP6 protein of strain FI-14 and subgroup II rotaviruses (Wa and Gottfried) indicates a high degree of homology (94%), while homology between strain FI-14 and subgroup I rotaviruses (SA-11, RF and 1076) was somewhat less (90 to 92%). On the other hand a high degree of conservation of amino acid sequence (95 to 97%) was observed between the H-2 strain and subgroup I rotaviruses. Five regions that may contribute to subgroup epitopes were identified. Region A (amino acids 45, 56) and region C (amino acids 114, 120) may contribute to subgroup I epitopes and regions B (amino acids 83, 86, 89, 92), D (amino acids 312 or 314, 317 or 319) and E (amino acids 341 or 343, 350 or 352) may contribute to subgroup II epitopes. When analysed using the Western blot technique monoclonal antibodies specific for VP6 epitopes shared by all rotaviruses were observed to react with both monomeric and trimeric forms of VP6, while monoclonal antibodies specific for a subgroup I or II epitope reacted only with the trimeric form of VP6. This observation and the sequence analyses suggest that subgroup antigenic specificity is determined by conformational epitopes produced by the folding of VP6 or the interaction between VP6 monomers.