Using a plaque assay with immunoglobulin (Ig)-coated SRBC, we and others have previously reported that the majority of polyclonally activated mouse lymphocytes secreted antibodies that appeared to be IgM anti-IgG autoantibodies. Careful reexamination of this assay, with application of several highly purified mouse serum and myeloma IgG and IgM preparations, revealed that IgM, which was a minor contaminant of Ig preparations, rather than IgG, was responsible for the formation of these plaques. High numbers of plaques could also be detected in assays with polyclonally activated human lymphocytes, Ig-coated SRBC, and anti-Ig developing sera. Of all IgG-, IgM- or IgA-secreting cells, 40 to 100% were detected with SRBC coated with gamma-globulin or Ig of the same isotype as the isotype to which the developing serum was specific; in general, low proportions of all PFC were detected with SRBC coated with Ig of a different isotype. Studies on the sequence of events leading to the formation of plaques with Ig-sensitized SRBC (both in humans and mice) revealed that antibodies detected in these assays were not able to bind to the Ig-coated SRBC (without the presence of developing serum), and therefore were not anti-Ig autoantibodies. It is our conclusion that the plaque assays with Ig-coated SRBC represent another type of a reverse hemolytic PFC assay that detects cells secreting antibodies regardless of their specificity, and these plaques are formed due to the cross-linking by the anti-Ig developing serum of the Ig coated on SRBC and the Ig secreted by lymphocytes. Our results confirmed preferential induction of anti-DNA antibody secreting cells in mice by showing that these antibodies indeed bind to DNA coated on SRBC. In cultures of polyclonally activated human lymphocytes, anti-DNA and anti-erythrocyte autoantibody-secreting cells were over 10 to 100 times less frequent than in mice. These results, therefore, disprove the concept of preferential induction of anti-Ig autoantibodies in the polyclonal activation of mouse and human lymphocytes, and show that anti-DNA and anti-erythrocyte autoantibodies are easily induced in the polyclonal activation of mouse, but not human, lymphocytes.