We argue that the proper way to treat thin-thick accretion-disk transitions should take into account the 2-D nature of the problem. We illustrate the physical inconsistency of the 1-D vertically integrated approach by discussing a particular example of the convective transport of energy.