A general circulation model is used to study the time evolution of a rotating, weakly baroclinic fluid in a basin with sloping sidewalls. Contours of f/h, where f is the Coriolis parameter and h is the depth of the fluid, are closed in this model. The fluid is forced by a localized source of positive vorticity. The initial response is a narrow, recirculating cell that resembles a β-plume modified by bathymetry. Such cells have been found in previous studies and have been linked to the recirculation cells observed in the subpolar North Atlantic. However, this is not a steady solution in this basin with closed f/h contours, and the circulation evolves into a gyre that encircles the basin. The time at which this transition occurs depends on the Rossby number, with higher Rossby numbers transitioning earlier. Based on the budget of potential vorticity, an argument is made that the western boundary is not long enough to drain significant vorticity from the flow and therefore a bathymetric β-plume is not a steady solution. A similar argument suggests that the Labrador Sea cannot sustain steady, linear, barotropic recirculations either. We speculate that the observed recirculations depend on inertial separation at sharp bathymetric gradients to break the assumption of linearity, which leads to significant viscous dissipation.