The experiments presented in this paper support the idea that the output of small lymphocytes from the thoracic duct of the rat (about 10⁹/day) is normally maintained by a large-scale re-circulation of cells from the blood to the lymph. It has been shown that the main channel from blood to lymph lies with in the lymph nodes and that small lymphocytes enter the nodes by crossing the walls of a specialized set of blood vessels, the post-capillary venules. In order to trace the fate of small lymphocytes, cells from the thoracic duct of rats were incubated for 1 h in vitro with tritiated adenosine. This labelled the RNA of about 65% of the small lymphocytes and more than 95% of the large lymphocytes; it also labelled the DNA of a proportion of the large lymphocytes. The mixture of small and large labelled lymphocytes was transfused into the blood of two groups of rats which belonged to the same highly inbred strain as the cell donors. At various times after the transfusions the thoracic ducts in one group of rats were cannulated to determine the proportion of labelled cells which could be recovered in the lymph; at corresponding times, the rats in the other group were killed and autoradiographs prepared from their tissues to determine the location of the labelled cells. The radioactive label in the RNA of small lymphocytes was stable enough to ensure that the labelled small lymphocytes which were recovered in the lymph several days after a transfusion were those which had originally been transfused into the blood. When the thoracic duct was cannulated 20 to 27 h after a transfusion, about 70% of the labelled small lymphocytes which had been transfused into the blood could be recovered from the thoracic duct over a 5-day period of lymph collection. During the first 36 to 48 h after cannulation, while the total output of small lymphocytes was falling rapidly, the proportion of labelled cells in the lymph remained approximately constant. The pool of the animal’s own cells with which the labelled cells had mixed contained between 1·5 and 2 × 10⁹ small lymphocytes; this was identified as the re-circulating pool. An autoradiographic study showed that after their transfusion into the blood the labelled small lymphocytes ‘homed’ rapidly and in large numbers into the lymph nodes, the white pulp of the spleen and the Peyer’s patches of the intestine. The concentration of labelled cells in other tissues was trivial in comparison. Labelled small lymphocytes were seen penetrating the endothelium of the post-capillary venules in the lymph nodes within 15 min of the start of a transfusion; they were traced into the cortex of the nodes and finally into the medullary lymph sinuses. Labelled small lymphocytes did not migrate into the adult thymus but a few entered the thymus of newborn rats. It was concluded that the re-circulating pool of small lymphocytes was located in the lymphoid tissue, the thymus excepted, and that the rapid ‘homing’ of cells into the lymph nodes had its basis in the special affinity of small lymphocytes for the endothelium of the post-capillary venules. The interpretation of these experiments was not complicated by the presence of large, as well as of small lymphocytes in the suspensions of labelled cells which were transfused. Other experiments, in which the large lymphocytes alone were labelled with tritiated thymidine, showed that most of them migrated from the blood into the wall of the gut where they assumed the appearance of primitive plasma cells; very few divided to form small lymphocytes.