The requirements for circumventing tolerance induction in favor of memory T cell development are poorly understood. Although two signals (Ag and costimulation) are necessary to drive effective T cell clonal expansion, few memory T cells remain after the response wanes. The adjuvant LPS can increase numbers of long-lived Ag-specific T cells, but its mechanism of action is not understood. In this report, it is shown that LPS, when combined with two-signal stimulation, profoundly enhances T cell survival in vivo. This survival does not appear to be dependent on the cytokines TNF-α, IL-1β, IL-6, and IFN-γ, nor is it dependent on the transcription factor NF-κB. However, in vivo proliferation of NF-κB-deficient T cells was comparable to that of wild-type T cells, yet their early accumulation in the lymph nodes was severely reduced unless the mice were treated with LPS and an agonistic CD40 mAb. Most importantly, we found that activation of two different costimulatory signals, CD40 and OX40, could not substitute for LPS in rescuing T cells from peripheral deletion. Perhaps surprisingly, these data show that LPS delivers a qualitatively different signal than multiple costimulatory signals.