Author Summary The class of immune cells called CD4 T lymphocytes consists of two major cell types: naïve cells that have not yet participated in an immune response and memory cells, which are cells that have responded to antigen, expanded in number, and acquired new characteristics. These two cell types can be distinguished from one another because they display different cell surface marker proteins. In this paper, we argue that many—probably most—of the cells researchers generally characterize as memory cells on the basis of their surface markers are not authentic memory cells. True memory cells—the ones produced, for example, when we immunize a child against a disease—divide very slowly, whereas the bulk of the cells we generally characterize as memory cells divide very rapidly. Mice that have never been exposed to antigens have as many of these “memory-like” cells as normal mice have, implying that these cells arise by a process that does not require foreign antigen. Analysis of the sequence of the antigen recognition receptors on these “memory-like” cells indicates that their replication does not derive from a few cells or clones undergoing multiple rounds of proliferation, thus their division cannot be explained by conventional, antigen-driven clonal expansion. We conclude that this large population of “memory-like” cells has arisen by a mechanism independent of a response to foreign antigen, and that these cells may have a crucial biological function.