Some previous work has suggested that fetal rejection may be a cause of spontaneous abortion in humans. The aim of the work presented, therefore, was to determine the influence of maternal alloantibody formation against paternal B lymphocytes, detected by the erythrocyte antibody rosette inhibition (EAI) assay, on the outcome of semi-allogeneic pregnancies. Preliminary studies indicated that EA rosette inhibition was a suitable assay for these investigations, because it detected alloantibodies directed to any major histocompatibility complex (MHC) antigen, irrespective of the ability of the antibody to fix complement; this was an important consideration because alloantibodies induced by pregnancy are often only weakly lytic. In humans it was found that antibodies to paternal B lymphocytes occurred significantly more commonly in normal primigravid and multigravid pregnancies when compared with pregnancies of similar gestation which aborted. These antibodies were shown to be directed to MHC encoded antigens by family studies, but were not removed by platelet absorption, strongly suggesting that they were not class I MHC antigens. Studies in inbred rats demonstrated that these paternal antigens were encoded by the RT1A region of the rat MHC alone. Maternal alloantibody responses to RT1A antigens appeared to be suppressive because studies using the rat kidney allograft model showed that multiparous rats with EAI antibodies to paternal strain cells enjoyed prolonged graft survival. It was also found that pregnancies in which the paternal strain differed only by RT1A antigens induced a suppressive immune response in the mother. These results suggest that immune responses to MHC encoded antigens, possibly unique, may prevent fetal rejection in some instances.