The two-dimensional version of the cumulus ensemble model developed by Soong and Ogura is applied both to a prestorm situation and to the mature stage of the extratropical mesoscale convective system (MCS) that developed on 10–11 April 1979 (AVE-SESAME-79 I) over the central United States. The objective is to investigate the statistical properties of convection, developing in response to an imposed large-scale forcing, and the thermodynamic feedback effect of clouds on the large-scale environment in midlatitudes. The result is compared to that recently obtained by Tao for a tropical rainband. The outstanding result of the model integration for 17 h of physical time is that statistical properties of clouds averaged horizontally over 128 km of the model domain undergo temporal variations for a given time-independent large-scale forcing, rather than settling down into a steady state. When applied to a prestorm situation, the model predicts heavy precipitation that continues to fall for the first 5 h, followed by a 4 h period without precipitation. A second burst of deep convection then occurs. An analysis of the result reveals that the pause of precipitation occurs when the subcloud layer is dried up primarily due to the net vertical transport of moisture associated with clouds. Convection again starts developing when the moisture in the subcloud layer is replenished by the imposed large-scale forcing. The precipitation rate averaged over the precipitation period is found to exceed the supply of moisture by the large-scale forcing. The result implies that the fraction of moisture convergence in a vertical air column that contributes to moisten the environmental atmosphere in Kuo's cumulus parameterization scheme can be negative. Further, the result indicates the following: 1) The updraft mass flux increases with height until it reaches the local maximum at 350 mb, indicating that the cloud population is dominated by deep clouds, in contrast to the bimodal or broad spectral distribution of clouds observed in the tropics. 2) The cloud heating effect does not balance the large-scale cooling effect, reflecting the fact that the storage and horizontal advection terms are not negligibly small compared to the vertical advection term in the large-scale heat budget; and 3) The net vertical fluxes of heat and moisture are not negligibly small cormpared to condensation and evaporation processes at upper levels in the heat and moisture budgets, reflecting the fact that the atmosphere considered here is more unstably stratified and updrafts are stronger than the tropical counterparts.