The constraint of fitting a diverse repertoire of antigen specificities in a limited total population of lymphocytes results in the frequency of naive cells specific for any given antigen (defined as the precursor frequency) being below the limit of detection by direct measurement. We have estimated this precursor frequency by titrating a known quantity of antigen-specific cells into naive recipients. Adoptive transfer of naive antigen-specific T cell receptor transgenic cells into syngeneic nontransgenic recipients, followed by stimulation with specific antigen, results in activation and expansion of both donor and endogenous antigen-specific cells in a dose-dependent manner. The precursor frequency is equal to the number of transferred cells when the transgenic and endogenous responses are of equal magnitude. Using this method we have estimated the precursor frequency of naive CD8 T cells specific for the H-2D^b–restricted GP33–41 epitope of LCMV to be 1 in 2 × 10⁵. Thus, in an uninfected mouse containing ∼2-4 × 10⁷ naive CD8 T cells we estimate there to be 100–200 epitope-specific cells. After LCMV infection these 100–200 GP33-specific naive CD8 T cells divide >14 times in 1 wk to reach a total of ∼10⁷ cells. Approximately 5% of these activated GP33-specific effector CD8 T cells survive to generate a memory pool consisting of ∼5 × 10⁵ cells. Thus, an acute LCMV infection results in a >1,000-fold increase in precursor frequency of D^bGP33-specific CD8 T cells from 2 × 10² naive cells in uninfected mice to 5 × 10⁵ memory cells in immunized mice.