SPAG-1, a Theileria annulata sporozoite surface antigen, is a vaccine candidate. Data is presented, based on the clonal segregation of SPAG-1 associated RFLPs, showing that this antigen is encoded by a single copy gene. We have cloned and sequenced a full-length genomic copy of the SPAG-1 gene and a comparison of this with a previously published SPAG-1 cDNA sequence demonstrates a high degree of polymorphism. We infer that these sequences represent two distinct allelic SPAG-1 variants. The deduced polypeptides show an overall identity of 92% with the most variable stretch (60% identity) occurring towards the middle of the molecule. The N and C termini are more conserved with identities of 92% and 97% respectively. The elastin receptor ligand, VGVAPG, present 3 times in the protein sequence derived from the cDNA is not found in that deduced from the genomic copy. Evidence for 2 further SPAG-1 alleles was obtained from PCR based sequences using macroschizont clones containing different SPAG-1 associated RFLPs. In summary we have shown the existence of at least 4 highly polymorphic SPAG-1 alleles. The implications of such polymorphism between and within distinct geographical isolates for the development of a SPAG-1 based subunit vaccine is discussed.