The classical cell-surface histocompatibility antigens (H–2 antigens in the mouse), known to have key roles in cell-to-cell recognition1, are encoded by at least three highly polymorphic genes (H–2D, K and L)2. Like their human (HLA) counterparts3, H–2 heavy chains span the cell membrane with a short C-terminal cytoplasmic region and an N-terminal extracellular stretch of about 280 amino acids. HLA antigens seem to be organized in three domains containing β-pleated sheets, with disulphide loops within the second and third domains, but the relative scarcity of material has hampered biochemical studies of the H–2 antigens4–6. We now report the sequencing of plasmids carrying H–2 cDNA as a means of inferring the amino acid sequence of the antigens, and especially of their previously poorly described C-terminal half.