Classical H-2 antigens can be divided into two, not very sharply defined groups, private and public . Private H-2 antigens (1) have a restricted strain distribution : among inbred strains they are found in only one haplotype of independent origin ; public H-2 antigens (2), on the other hand, are shared by at least two, and quite often by a large number of unrelated haplotypes . Private H-2 antigens are believed to be controlled by alleles of two loci at the opposite ends of the H-2 complex, the H-2K locus in the K region, and the H-2D locus in the D region . In agreement with this belief, the antigens can be arranged in two series, where members of each series are mutually exclusive (3) . No such arrangement is possible for the public H-2 antigens and the nature of the genetic control of these antigens has long been a matter of controversy .According to the traditional interpretation of the H-2 complex (4), the public antigens are controlled by separate regions located between the K and D regions within the complex . According to the two-locus model (5), public H-2 antigens reside in the same molecules as the private ones and are controlled either by the H-2K or H-2D loci, or both . The two-locus model fits better the available data than the traditional multi-locus model (6) . Also, the former model is supported by the results of transplantation analysis demonstrating that only the peripheral regions of the H-2 complex are involved in rapid skin-graft rejection (7) . However, a formal proof that the serologically detectable public H-2 antigens are coded for by the K and D regions has never been provided . As a matter of fact, most recent developments (8, 9, cf. Discussion) suggest that an across the board assignment of public antigens to K and D regions could not be justified . Despite these developments, many investigators accept the two-locus model as an established fact and rely on it in their interpretation of the H-2 system . At this stage it is, therefore, critical to test whether the serologically detectable public H-2 antigens conform to the predictions of the two-locus model .We have recently developed a method that can be used to examine molecular relationships among various cell membrane antigens . The method is based on the observation that under certain circumstances, antigens coated with alloanti-* This investigation was supported by grant