There is renewed interest in the processes responsible for the formation of the oceanic main thermocline and deep circulation, but analytical techniques remain inadequate for the study of even idealized thermocline systems in closed basins because of their nonlinear boundary currents, mixing, and dissipation. Most closed basin work has been done with quasi-geostrophic (QG) model physics, but the QG system has limitations, particularly for studies of the deep circulation. It is subject to several lower layer vorticity-forcing and dissipation constraints that have no clear primitive equation counterparts. These constraints are likely to introduce both special deep flow characteristics and considerable sensitivity to the type and magnitude of lower layer friction assumed. As a first look at some of the effects of these constraints, we have carried out (1) a series of QG experiments that can be compared with otherwise identical primitive equation (PE) experiments and (2) a second set of QG experiments in which the lower layer friction is changed in form. In every otherwise identical PE/QG comparison, significant quantitative and/or qualitative changes are found in the deep circulation. Nontrivial upper layer flow changes are sometimes also found. Deep layer friction changes in the QG systems also induce substantial deep flow changes. These results suggest that the limitations and sensitivity of QG physics make it less than ideal for model studies of deep mean circulation physics and hence for detailed tuning against deep ocean data. It is important that the behavior of PE systems be examined with new calculations if our understanding of the deep circulation is to advance.