The results discussed here provide strong evidence that different T-cell effector gene programs are activated by different signals, and that in several cases their responses to the same exogenous stimuli shift during the development and antigen responses of the cells. T-cell responses are thus conditional and plastic at the individual cell level. In the formalism of the introductory section, the results support elements of Models 2 and 3, and suggest a fusion between them as differentiation is explained in terms of alteration in the relative strengths of different intracellular signaling pathways. Returning to an initial question, how are different functional capabilities assigned nonrandomly to cells with different antigen recognition specificities? This question has not been answered, but it can be reformulated. If all virgin T cells can transiently make IL-2, then we must ask what features of cell biology explain the preferential preservation of IL-2 inducibility in CD4+ cells as opposed to CD8+ cells. If the capacity to induce IL-4 expression is not acquired in the thymus, then we may ask whether the initial opening of this locus depends on a CD4-transmitted signal. Similarly, the CD8 molecule itself might participate in inducing the initial differentiation events that render CTL-p inducible for granzyme C and perforin. This would be in accord with a large literature showing that CD8 engagement is much more important in the initial induction of CTL activity than in the exercise of function by pre-primed CTL effectors. The subtext of each of these "questions", however, is that intrathymic events may not directly affect the genes used by terminal effectors for function at all. They may instead bias a cell's complement of triggering receptors, thus rendering it differentially sensitive to particular signals generated during antigen reception. This view is extreme, and will probably turn out to be an overstatement. But it does inspire a unique set of investigations into the basis of T-cell function. It lends urgency to the question of whether CD4+ and CD8+ cells differ in their G proteins, kinases, or inducible proto-oncogenes. If they do, we can then ask whether such differences themselves arise in the periphery, or whether they can be traced back to thymocytes fresh from positive selection--or before.