Observations of air flow in and around convective clouds are summarized and discussed in light of the requirements for parameterization of midlatitude convection. Both theory and observation indicate that a substantial portion of the compensating subsidence occurs as dry downdrafts in the immediate vicinity of convective clouds, which suggests that an additional physical mechanism is necessary in linking convection to the large scale. A conceptual three-dimensional model of midlatitude organized convection along with its implications for parameterization, particularly in mesoscale modeling, is presented. The effects of moist downdrafts are also considered and some differences between midlatitude and tropical convection are discussed.