To analyze the critical parameters for effective antiviral cytotoxic T lymphocyte (CTL) activity in vivo, control of lymphocytic choriomeningitis virus (LCMV) infection in the spleen was studied after adoptive transfer of different spleen cell populations into preinfected recipients. The quantitative, qualitative and kinetic requirements for virus control were defined and related to in vitro assays to compare the antiviral protective function of CTL from naive, acutely infected and memory mice. Treatment of mice with an established but limited LCMV infection by adoptive transfer of spleen cells from acutely LCMV-infected mice led to complete virus elimination mainly mediated by donor-derived CD8+ T cell-mediated, perforin-dependent cytotoxicity. Since virus is continuously spreading and the number of infected target cells rapidly increases, the time until target cell lysis is achieved was critical: if release of viral progeny was not prevented early, additional time to perform effector function did not improve overall virus control. When the function of various cell populations was compared in this model, we found that CTL from naive and memory mice perform considerably less well than CTL from acutely infected mice. In vitro studies indicated that this is probably due to the fact that they can not fulfill the limiting time requirements for immediate antiviral protection: while CTL from acutely infected mice can perform lytic effector function immediately, memory CTL require a considerable reactivation time before they can lyse infected target cells. This reactivation does not necessarily involve cell division. These findings illustrate how critical time limitations are for CTL to mediate early control of a dynamic virus infection in vivo.