A patient, O.A., had in her serum an IgM rheumatoid factor (RF) of high titer which reacted equally well with monomeric and polymeric rabbit IgG, but showed no reaction with homologous or autologous monomeric and polymeric IgG, neither in hemagglutination nor in ELISA tests. In contrast to the usual specificity of RFs crossreacting with rabbit and human IgG, which require the presence of both the CH2 end CH3 domains, the O.A. RF showed reaction only against the CH3 domain. The existence of human RF not reacting with human IgG could challenge the notion that RF production is necessarily induced by autologous IgG and thus playing an important role in potentiating and regulating secondary immune responses. The O.A. RF might represent 1: an extreme case of heteroclitic antibody, 2: the internal image of a hypothetical Fc-receptor reacting with rabbit IgG but not with human IgG, 3: a result of polyclonal B cell activation, 4: an antigenic crossreaction with an unknown antigen with similarities to rabbit IgG but not to human IgG.