AbstractWhen normal human B cells from blood, tonsil or spleen are cultured with 20–30% autologous or allogeneic T cells and mitogen, the majority of cells rosette with sheep erythrocytes (E) after 3–5 days in culture. These E+cells are of transformed appearance and several points indicate that many are derived from B cells. Some simultaneously possess surface immunoglobulin (sIg) or a B cell antigen [detected with BA1 monoclonal antibody (mAb)] and E receptor. Many do not stain with anti‐T cell mAb (UCHT1, OKT4, OKT8), while cultured T cells continue to express these antigens. Furthermore, many E+cells appear when the original T cells have been irradiated. All the cultured E+cells stain with OKT11, an mAb against the E receptor, and their positivity cannot therefore be attributed to mitogen‐induced nonspecific stickiness. The E positivity of the B cells was shown to be endogenous since passive acquisition of E receptor shed by T cells was excluded in a number of ways; no phenotypic changes were observed when supernatants from cultures containing many E+sIg+non‐T cells were added to B cells, or when the B and T cells were separated by a Millipore membrane; and E receptor was re‐expressed after stripping with trypsin or pronase. The intimate presence of T cells is essential for the expression of E receptors by B cells, and this helper capacity was shown to reside within the OKT4‐defined helper T cell subpopulation. The significance of the expression of E receptor by B cells is discussed in relation to the recentin vitroandin vivodemonstration of E+sIg+cells in certain leukemias and in relation to the specificity of E rosetting as a mark