The reduced immunological responsiveness of neonatally thymectomized rodents has been reversed far more effectively with cells from spleen, lymph nodes, and thoracic-duct lymph than with cells from either thymus or bone marrow.' In heavily irradiated mice, an inoculum containing a mixture of marrow and thymus cells allowed the production of more hemolysins against sheep erythrocytes than could be accounted for by summating the activities of each cell population alone.2 Some type of interaction must thus take place between thymus cells, bone marrow cells, and antigen. In order to characterize the nature of this interaction one would first have to determine which cell type, thymus or bone marrow, was the immediate precursor of the antibody-forming cell.In thymectomized irradiated mice protected with chromosomally marked bone marrow and grafted with chromosomally marked thymus tissue, both marrowdonor-type and thymus-donor-type cells were found dividing in the lymphoid tissues. The majority of dividing cells were of marrow donor origin,3 but a sharp increase in the proportion of thymus-derived cells occurred for a short period of time after antigenic stimulation.4 An attempt was made to determine the origin of the antibody-forming cell by using a transfer system in which both chromosome and isoantigenic markers were available. The results indicated that, although thymus-graft derived cells responded vigorously to antigenic stimulation by mitosis, the antibody-forming cells had the immunogenetic characteristics of the marrow donor.4 However, it is known that the lymphocyte population of a thymus graft in a marrow-protected irradiated host is entirely replaced within two to three weeks by cells from the inoculated marrow.3 Hence, although antibody-forming cells had the genotypic characteristics of the marrow donor, they could have been derived from cells that had first migrated through the thymus implant which, in these experiments, had been grafted 30 days prior to antigenic stimulation. Such an experimental design therefore failed to establish unequivocally which cell type was the immediate precursor of the antibody-forming cell: thymus or marrow cell.In our laboratory we have attacked the problem by using cell suspensions rather than thymus grafts. A mixture of thymus and bone marrow cells from unrelated strains of mice was inoculated together with sheep erythrocytes into irradiated hosts. Under these conditions no hemolysin response was obtained, I suggesting that interaction does not take place between allogeneic cells immediately upon transfer to irradiated recipients. Neonatally thymectomized mice were then used as hosts of syngeneic or allogeneic thymus cells injected together with sheep erythrocytes. Since the marrow of neonatally thymectomized rodents 296