A wide body of modeling and theoretical scaling studies support the concept that changes to the Atlantic meridional overturning circulation (AMOC), whether forced by winds or buoyancy fluxes, can be understood in terms of a simple causative relation between the AMOC and an appropriately defined meridional density gradient (MDG). The MDG is supposed to translate directly into a meridional pressure gradient. Here two sets of experiments are performed using a modular ocean model coupled to an energy-moisture balance model in which the positive AMOC-MDGrelation breaks down. In the first suite of seven model integrations it is found that increasing winds in the Southern Ocean cause an increase in overturning while the surface density difference between the equator and North Atlantic drops. In the second suite of eight model integrations the equation of state is manipulated so that the density is calculated at the model temperature plus an artificial increment ?T that ranges from-3° to 9°C. (An increase in ?T results in increased sensitivity of density to temperature gradients.) The AMOC in these model integrations drops as the MDG increases regardless of whether the density difference is computed at the surface or averaged over the upper ocean. Traditional scaling analysis can only produce this weaker AMOC if the scale depth decreases enough to compensate for the stronger MDG. Five estimates of the depth scale are evaluated and it is found that the changes in theAMOCcan be derived from scaling analysis when using the depth of the maximumoverturning circulation or estimates thereof but not from the pycnocline depth. These two depth scales are commonly assumed to be the same in theoretical models of the AMOC. It is suggested that the correlation between the MDG and AMOC breaks down in these model integrations because the depth and strength of the AMOC is influenced strongly by remote forcing such as Southern Ocean winds and Antarctic Bottom Water formation.