H-2 Specificities 23, 25, 11, 1, and 5, when arranged in this order, form an inclusion or subtype system, with 23 showing the narrowest distribution. Three of these specificities, 23, 25, and 11, are assigned to the K crossover region; the other two usually are assigned to the E region. Because of ambiguities in regard to 1 and 5, these specificities were selected for detailed study. Our panel of strains included two, PL/J and SM/J, with new alleles designated, respectively,H-2uandH-2v. Included also was strain B10. Y, previously typed asH-2q, but which now has an allele similar to or identical withH-2p. We provisionally call itpa.The panel did not includenin terms of whichH-2.1was originally defined. However, allelespandnare similar, and our anti-pand anti-pasera contained antibodies with reactions closely approximating those expected of an anti-1. Recipients for the production of antisera were selected to exclude as many antibodies as possible other than 1. Antisera were analyzed by direct hemagglutination and cytotoxicity against standard panels of cells. Many were also analyzed by in vivo hemagglutination absorption, and some, by quantitative in vitro cytotoxic absorption. Antisera made against allelesp, pa, s, v, andowere 1-like, in the range of their reactions, but showed individual differences. An antiserum made against alleleqalso contained a 1-like antibody, although it was complicated by the presence of anti-11. Antisera made withpas one parent of the recipient and withsas the donor (and also in the reverse combination) contained 1-like antibodies. We refer to these as blocked anti-1 sera. All unblocked anti-por anti-pasera reacted by hemagglutination withs, and one was fully absorbed bys.The individual differences in the reaction patterns of the antisera were largely confined to the 1-positive, 23-negative alleles. We refer to these as 1-intermediate alleles. Reactions with these alleles were often weak. With the 1-positive, 23-positive alleles (designated 1-complete alleles), except for some weak reactions by the blocked sera, the reactions were consistently strong. The differences in strength of reactivity between 1-intermediate and 1-complete alleles were particularly clear in the cytotoxic absorption tests. We suggest as a possible explanation a unidirectional cross reaction between 1 and 23. Crossover alleleoreceived 1, but not 23, from 1-complete allelek; as would be predicted, it behaved like a 1-intermediate. We suggest that the variability of the reactions of the 1-intermediate alleles reflects mutational variation in their 1 sites, sufficient similarity remaining so that cross reactions occur, however. This hypothesis also explains the formation of 1-like antibodies in blocked combinations. The 1-complete alleles, because their EK regions trace through successive