Three separate forms of in vitro stimulation were assessed for their capacity to activate hapten-gelatin fractionated, fluorescein- (Flu) specific murine splenic B lymphocytes. They were: a) Flu-polymerized flagellin (Flu-POL) acting on a single Flu-specific B cell in microculture in the absence of "filler" or feeder cells, but in the presence of T cell-derived B cell growth and differentiation factor(s) (BGDF); b) a mixture of mitogens, E. coli lipopolysaccharide (LPS) and dextran sulfate, acting on a single Flu-specific B cell in the absence of added BGDF; and c) Flu-POL plus BGDF acting on single Flu-specific B cells as in a but with thymus filler cells also present. System c was markedly superior in causing antibody formation, 15 to 22% of cells forming a clone of Flu-specific antibody-forming cells (AFC), in contrast to 6% for system b and 3 to 6% for system a. Each stimulus was applied to single cells that had been size fractionated into samples of increasing size by using the forward light-scattering parameter of the fluorescence-activated cell sorter. Surprisingly, the smaller sized fractions proliferated poorly in system a and contributed less than 10% of the antibody-forming potential of the total population. The smaller cells proliferated better in system b, but only 10 to 15% of proliferating clones generated Flu-specific AFC, whereas the larger cells contributed 86% of the total AFC response. Even in system c, only 6% of the small cells formed AFC clones compared with 41% of the larger cells. It thus appears that the smaller half of murine B lymphocytes is relatively resistant to activation into proliferation and differentiation by "T-independent" antigens; when activated by mitogens, they clone less efficiently than larger cells. Despite these limitations, system c could generate a total of up to four hapten-specific AFC for every B cell placed into culture, making it the most efficient system of specific antibody formation yet described.