Ureaplasma urealyticum has been subspe-ciated by a number of serologic methods. Eight serotypes have been identified by modified metabolic inhibition, growth inhibition and indirect hemagglutination. Fourteen serovars have been identified by immunofluorescence and 16 by the mycoplasmacidal assay. The present studies were performed to determine if group-specific antigens could be detected by immunofluorescence and if group- or serovar-specific antigens could be detected by enzyme-linked immunosorbent assay and immunoblotting. Reaction of rabbit antisera to U. urealyticum with homologous and heterologous serotypes based upon end point immunofluorescent titration did not differentiate the serovars into the two biotypes. In attempts to quantitate the number of a given serovar present in clinical specimens, the total number of colonies that were stained often exceeded 100%, suggesting either that the serovars represented in stock cultures are not truly representative of those present in humans or that certain strains express multiple serovar specificities. End point titration in an enzyme-linked immunoadsorbent assay using either whole cells or cell lysates and rabbit antisera failed to distinguish between group and serovars (i.e. in many cases end point titers were only 2-fold lower with heterologous antigens). Furthermore convalescent sera from patients infected with a single serovar failed to induce a serovar-specific response detectable in an enzyme-linked immunosorbent assay. Likewise immunoblotting of one dimensional electropho-retograms with sera from humans known to be infected with a single serovar showed that antibodies against that serovar recognized peptides present in all serovars and reacted with those bands with various degrees of intensity. Thus immunoblotting with human sera also failed to reveal either group- or serovar-specific polypeptides. Although immunoblotting with rabbit antisera revealed a certain degree of relatedness between members of a group, a unique group- or serovar-specific protein could not be demonstrated. Thus it seems likely that group-and serovar-specificities may be due to quantitative rather than qualitative differences or that nonprotein antigens are involved.