This study compares the functional properties of rat thoracic duct lymphocytes (TDL) after stimulation with strong alloantigens of the major histocompatibility complex (MHC) either in vitro in preparative mixed lymphocyte interactions (MLI) or in vivo in systemic graft-vs-host (GVH) reactions. Comparisons were made of PHA responses and reactivity to the specific priming haplotypes or to third party haplotypes in analytical MLI and in GVH reactions either before or after the activated populations were "parked" in syngenetic T cell-deprived (B) rats. These comparisons can be summarized as follows: 1) TDL populations primed in bulk MLI cultures (MLI-TDL) slowed some evidence of specific positive selection when tested immediately; MLI responses to specific alloantigens were both relatively large and accelerated in tempo, whereas responses to third party alloantigens were diminished but also accelerated in tempo. Specific GVH responses were more marked than in third party recipients but they were also decreased relative to normal, and displayed an abberant dose/response slope. MLI-TDL populations tested after they had been stored in syngeneic B rats showed clear evidence of stable-specific positive selection; specific MLI and GVH responses were enriched relative to third party responses and also in comparison to normal, unselected TDL populations. This finding indicates that GVH and MLI reactivity are probably both functional capacities of the same lymphocyte subpopulation since positive selection by one function (MLI) also enriched for a second (GVH). 2) Parental strain TDL activated in vivo in the systemic GVH reaction in irradiated F1 animals and recovered from the thoracic duct 3 to 4 days later (late GVH-TDL) consisted mainly of blast cells, however, in contrast to MLI-TDL these populations showed no evidence of positive selection when tested before or after parking in B rats. MLI responses to specific alloantigens were minimal, and greatly reduced in magnitude compared to normal. GVH responses to specific haplotypes could be detected, but these were not enriched compared to normal, despite the content in the late GVH-TDL populations of a significant proportion of blast cells presumably activated by host alloantigens. 3) Early collections (less than 40 hr) of parental strain GVH-TDL collected from F1 recipients contained no blast cells and showed impressive degrees of negative selection; they were markedly depleted of both GVH and MLI activity to specific alloantigens but displayed normal reactivity to third party alloantigens. Moreover, specific negative selection was persistent in these populations parked for several weeks in B rats, and indication that a specific subpopulation of reactive cells had been physically eliminated. 4) PHA responses of both MLI- and GVH-activated TDL populations tested either before or after parking in B rats were approximately normal on a per T cell basis...