Rabbits were immunized with the sera from patients with known Au antigenaemia; from kidney homograft recipients who were Au negative with direct conventional tests but who had hepatic dysfunction; and from normal volunteers. Heterologous anti-Au antibodies were raised by the first of these kinds of sera but not by the third. Sera from seven of the eleven immunosuppressed kidney recipients of the second group did not raise anti-Au antibodies in the rabbit, but sera from the other four did. The results indicate that the Au antigen may be present in trace quantities in significant numbers of immunosuppressed patients previously thought to be Au negative. However, they do not unequivocally establish an aetiologic association between Au antigenaemia and `post-transplantation liver disease', probably because of the obscuring and important factor of hepatotoxicity of the immunosuppressive agents.