Abstract Vaccines that elicit T cell responses try to mimic protective memory T cell immunity after infection by increasing the frequency of Ag-specific T cells in the immune repertoire. However, the factors that determine immunodominance during infection and after vaccination and the relation between immunodominance and protection are incompletely understood. We previously identified TB10.4(20–28) as an immunodominant epitope recognized by H2-Kd–restricted CD8+ T cells after M. tuberculosis infection. Here we report a second epitope, EspA(150–158), that is recognized by a substantial number of pulmonary CD8+ T cells. The relative abundance of these T cells in the naive repertoire only partially predicts their relative frequency after M. tuberculosis infection. Furthermore, although vaccination with recombinant vaccinia virus expressing these epitopes changes their relative immunodominance in the preinfection T cell repertoire, this change is transient after challenge with M. tuberculosis. We speculate that factors intrinsic to the chronic nature of M. tuberculosis infection establishes the hierarchy of immunodominance and may explain the failure of some vaccines to provide protection.