Most tumors grow progressively and overwhelm the host. The rare but documented cases of spontaneous regression of primary tumors are indicative of the potential of tumor-bearing hosts to develop a significant antitumor response. Because most tumors grow progressively in the host, it is not surprising that the majority of studies have focused on T lymphocytes that infiltrate these tumors. Although these studies have generated significant and useful information during the period of tumor growth, they can only speculate on the mechanisms that are involved in tumor rejection. We have used a well developed sponge model of concomitant tumor immunity that allows us to compare the immunologic events that occur during tumor progression vs rejection. In this model, an animal harboring a primary EMT6 mammary tumor is challenged with a secondary tumor implant through a pre-implanted gelatin sponge. During the manifestation of concomitant tumor immunity, the secondary tumor is rejected and the effector cells mediating the response are retained within the sponge matrix. Using this model we analyzed the TCR usage, cytotoxic activity of lymphocytes, and cytokine production at both tumor sites. The data revealed that tumor-rejecting lymphocytes isolated from the site of secondary tumor implant were cytotoxic toward EMT6 cells, whereas tumor-infiltrating lymphocytes isolated from the progressing primary tumor were not. Interestingly, the TCR-V beta repertoire of the tumor-infiltrating lymphocytes and tumor-rejecting lymphocytes were identical with V beta 1 and V beta 8 being predominant at both sites. Furthermore, the rejection site showed higher gene expression of IFN-gamma, TNF-alpha, and IL-10 whereas TGF-beta expression was slightly higher in the progressing tumors. These findings suggest that the disparate effector functions observed during tumor progression vs rejection are not caused by different T cell phenotypes but may be due instead to influences exerted by cytokines produced at the tumor sites.