The CD8+ T cell response to the immunodominant DbNP366 epitope has been analyzed sequentially to determine the prevalence and persistence of different T cell antigen receptor (TCR)Vbeta8.3 clonotypes after primary and secondary influenza virus challenge. Based on the length and amino acid sequences of the complementarity-determining region 3 of TCRbeta (CDR3beta) loop and associated Jbeta usage, the same dominant TCRbeta signatures were found in the blood, the spleen, and the site of virus-induced pathology in the infected respiratory tract. Longitudinal analysis demonstrated that TCRbeta prominent in the antigen-driven phase of response persisted into memory and were again expanded after secondary challenge. A proportion of these high-frequency TCRbeta expressed "public" CDR3beta sequences that were detected in every mouse sampled, whereas others were found more than once but were not invariably present. Analysis of N-region nucleotide diversity established that as many as 10 different nucleic acid sequences (maximum of four "nucleotypes" in any one mouse) could encode a single public TCRbeta amino acid sequence. Conversely, whereas some of the unique, "private" TCRbeta achieved a substantial clone size, they were always specified by a single nucleotype. Although there is a strong stochastic element in this response, the public TCRbeta seem to represent a "best fit" for this immunodominant epitope, are selected preferentially from the naive TCR repertoire, and assume even greater prominence after secondary challenge.