A detailed understanding of the effects of costimulatory signals on primary T cell expansion has been limited by experimental approaches that measure the bulk response of a cell population, without distinguishing responses of individual cells. Here, we have labeled live T cells in vitro with a stable, fluorescent dye that segregates equally between daughter cells upon cell division, allowing the proliferative history of any T cell present or generated during a response to be monitored over time. This system permits simultaneous evaluation of T cell surface markers, allowing concomitant assessment of cellular activation and quantitative determination of T cell receptor (TCR) occupancy on individual cells. Through this approach, we find that TCR engagement primarily regulates the frequency of T cells that enter the proliferative pool, but has relatively little effect on the number of times these cells will ultimately divide. In contrast, CD28-costimulation regulates both the frequency of responding cells (particularly at sub-maximal levels of TCR engagement), and more prominently, the number of mitotic events that responding cells undergo. When CD28-stimulation is blocked, provision of IL-2 restores the frequency of responding cells and the normal pattern of mitotic progression, indicating that the other CD28-induced genes are not required for this effect. An unexpected finding was that even at maximal levels of TCR engagement and CD28-mediated costimulation, only 50-60% of the original T cells in culture can be induced to divide. The nondividing cells are heterogeneous for naive versus memory markers, suggesting a more complex relationship between expression of memory markers and the ability to be recruited into the dividing pool. From these studies, we conclude that a stringent checkpoint regulates the participation of activated T cells in clonal expansion, with TCR and CD28 signals having both overlapping and differential effects on the induction and maintenance of T cell responses.