Summary: We are often taught that secondary, or memory, responses by lymphocytes are more vigorous than primary responses. An expectation commonly associated with this notion is that the initial encounter with a pathogen should result in immunity to re-infection. Although this outcome is sometimes the case, it is not universally true. In this review, we propose a unified model of T-cell memory to explain the apparent successes and failures of eliciting vaccine-like protection from prior encounters with pathogens. We speculate that memory T cells arise as an invariant consequence of clonal selection during an immune response. The quality of memory T cells, however, seems to vary in the degree to which they have acquired effector characteristics and, thus, their ability to confer immunity to re-infection. Although not all memory T cells possess the embellished attributes of fully developed effector cells, they all seem to share the rudimentary quality of preserving an antigen specificity that has proven itself useful. We suggest that the ability to maintain the integrity of the T-cell repertoire, more than establishing immunity to re-infection, may represent the fundamental form of memory for the adaptive immune system.