Spleen cells from non-immunized adult mice were fractionated on thin layers of fluorescein (FLU)-gelatin to yield FLU-specific B cells. These B cells were cultured either singly or in very small numbers in 10 microliter microcultures with 0.1 microgram/ml of the T cell-independent (TI) antigen FLU-E. coli lipopolysaccharide (FLU-LPS). Cultures were either filler cell-free, or supported by the addition of 10(5) CBA/N thymus cells per well. At 4-6 days, culture supernatants were assayed for the presence of anti-FLU antibody either by an enzyme-linked immunosorbent assay (ELISA) or a radioimmunoassay (RIA). With the filler cell-free cultures, B cell proliferation was scored microscopically before removal of culture supernatant. The cultured cells from each well were assayed for their capacity to form directly hemolytic FLU-specific plaques. In the filler cell-free system, the ELISA was much more sensitive than the plaque assay in identifying antibody-forming cell (AFC) clones, with over 10% of fractionated B cells yielding clones secreting detectable antibody, though with a low mean optical density (OD). This value represented over 80% of the proliferating clones. In the more efficient, filler cell-supported system, the difference between the 2 read-out methods was smaller. Here, one half of the hapten-specific B cells formed AFC clones, the highest cloning efficiency yet reported for an antigen-driven system. Comparative studies showed the RIA to be only marginally more sensitive than the ELISA, and not nearly as convenient for routine use.