To study the development of alloreactive cytolytic T cells in vivo, C57BL/6 mice were implanted s.c. with polyurethane sponges bearing allogeneic (DBA/2) splenocytes. On various days thereafter, cells that had accumulated in these sponge grafts were tested for cytolytic activity against DBA/2 target cells in 51Cr-release assays, and for frequency of DBA/2-reactive CTL as determined by limiting dilution analysis (LDA). During these studies we found that LDA was consistently more efficient at detecting alloreactive CTL than the traditional 51Cr-release assays. As determined by LDA, sponge grafts initially infused with DBA/2 splenocytes acquired high levels of DBA/2-reactive CTL, while sponge grafts infused only with saline acquired few DBA/2-reactive CTL. DBA/2-reactive CTL first became detectable in sponge allografts approximately four days after implantation, and reached a maximal frequency by the 10th day after implantation. This frequency was maintained for at least the next seven days. In contrast, the ability of cellular infiltrates from sponge allografts to lyse DBA/2 target cells in 51Cr-release assays was not detectable until the 7th day after implantation, was optimal by the tenth day, but declined thereafter to lower levels, as observed on the 13th and 17th day after implantation. Since the frequency of CTL remains stable through the 17th day after implantation, this decline in cytolytic activity may indicate that donor-reactive CTL remain in sponge allografts, but they continue to differentiate to a noncytolytic status. We further observed that previous allosensitization with skin grafts markedly accelerates the accumulation of alloreactive CTL in sponge allografts. The mechanism that promotes more rapid accumulation of CTL in allosensitized sponge graft recipients remains to be established. Throughout these studies, we observed that even at peak development, donor reactive CTL are, at most, 0.2% of the cells recovered from sponge allografts. This raises some questions regarding not only the fundamental role of the CTL in allograft rejection, but also the role of the remaining 99.8% of the allograft-infiltrating cells.