When initial antigen encounter involves optimal antigenic and costimulatory stimuli, naïve CD8 T cells undergo a developmental program that leads to their activation, expansion and acquisition of effector functions (including production of IL-2, IFNgamma and expression of cytolytic effector molecules). A subset of the activated CD8 T cells thrives as long-lived memory cells. Encounter of tissue-associated, and in particular tumor-associated antigen, may often be suboptimal in terms of antigenicity and costimulation, however. We previously developed a model of naïve CD8 T cells from transgenic mice expressing an alloreactive TCR for which a mutant alloantigen behaved as a partial agonist, inducing only some of the effector functions induced by the native alloantigen. To ascertain the molecular bases for the establishment of divergent fates within the same naïve CD8 T cells, we have used cDNA microarrays to monitor sequential gene expression patterns in conditions of full or partial response of these naïve CD8 T cells. Of the 5000 different genes monitored on the array, 18% showed changes in expression in activated versus naïve CD8 T cells, independent of whether stimulation was with full or partial agonist. These included antigen-induced upregulated as well as downregulated genes. Clusters of genes that were differentially expressed were also identified, being either (i) weakly versus strongly, or (ii) transiently versus stably expressed in response to partial and full agonist, respectively. They included (i) genes encoding costimulatory molecules and (ii) genes controlling cytolytic function, cytokine production, and chemokines. Therefore, the cDNA microarray approach was a sensitive tool to provide an exhaustive picture of T cell activation as it could discriminate quantitative, qualitative and dynamic differences in mRNA expression profiles between fully or partially activated T cells.