The authors discuss laboratory experiments that elucidate the mechanism of formation and westward drift of anticyclonic baroclinic vortices from a buoyant surface current flowing along a lateral boundary and around a cape. Experiments were carried out with a sloping bottom in order to simulate the topographic β effect. They showed how a vortex can be generated from the current where it separates and reattaches to the cape and that, under some conditions, the eddy is able to detach from the cape and drift westward following isobaths. Two important timescales regulate the flow: the time tf that the current takes to generate a vortex and the time td that the vortex takes to drift westward for a distance equal to its radius. When these two timescales are either of the same order of magnitude or tf < td, the eddy was observed to translate westward. For tf > td the vortex was able to form at the cape but it did not detach and drift westward. The influence of the depth of the lower layer, h0, on the flow was investigated. The theoretical westward speed U2d depends on the depth of the lower layer:the deeper the lower layer the slower the drift. The values of the slope s required in the experiments in order to obtain the detachment and drift of the vortex indicate that the phenomena will occur on a planetary β plane only when the variation of the Coriolis parameter with latitude is reinforced by a topographic β effect. A good agreement between the laboratory experiments and the observations of meddies in the Canary Basin, where the Mediterranean Outflow from the Strait of Gibraltar flows along the coast of Spain and around Cape St. Vincent, suggests that the eddy-shedding process is similar to that observed in the laboratory experiments.