Two different H deficient phenotypes are observed in one kindred. Three of them in two generations of a same family appear as Bombay like. In the other branch of the kindred, an Hz phenotype (as described in the editorial of this issue) is observed. The most simple explanation is that the three Bombay like phenotypes correspond to Hz ABH non secretor (Hz sese) individuals, this being indistinguishable from a true Bombay. The high level of I antigen in the plasma of the three Bombay like (as observed in Hz in contrast to true Bombay) could favour such an hypothesis. According to ORIOL'S new hypothesis [5], III2, III3 and IV2 would genetically be hh sese, III8 would be hh Se and the h Se would therefore be a recombining haplotype (the original haplotype being h se).