T cells potentially encounter numerous endogenous peptides during selection in the thymus and in the periphery. We examined the impact of an endogenous peptide on in vivo T cell development, using a TCR transgenic mouse model based on a hemoglobin-specific T cell clone. In these mice, the transgenic -chains paired with endogenous ot chains. This led to a serendipitous primary reactivity to Set69 peptide, an altered peptide ligand of the Hb a (64-76) epitope of the parent clone. Two Ser69-reactive T cell populations were identified. A smaller population of the Ser69-reactive T cells responded both to Set69 and Hb a (64-76). A majority reacted only to Set69, and not to Hbd(64-76); in fact, Hbd(64-76) was a specific TCR antagonist for these Ser69-onlyreactive T cells. Thus, in this unique experimental system, Ser69 became an agonist, and Hbd (64-76) was an antagonist. Endogenous presentation of the antagonist ligand in the thymus selectively eliminated the high-avidity cells, while sparing low-avidity cells in the Ser69-reactive T cell repertoire. These results highlight how specificity guides developing T cells through a network of ligands and indicate that the endogenous peptide pool has a profound effect on T cell development and repertoire.