The present study investigates the interaction between a self-propagating cyclonic eddy with two right vertical cylinders and determines the conditions for an eddy to bifurcate into two or more eddies. As in previous studies, after the cyclonic eddy came in contact with a cylinder, fluid peeled off the outer edge of the vortex and a so-called “streamer” went around the cylinder in a counterclockwise direction. Under the right conditions, this fluid formed a new cyclonic vortex in the wake of the cylinder, causing bifurcation of the original vortex into two vortices. In some cases, two “streamers” formed and went around the two cylinders, each forming a new cyclonic vortex. During the experiments, three parameters were varied: G, the separation between the cylinders; d, the diameter of the incident vortex; and y, the distance of the center of the vortex from an axis passing through the center of the gap between the cylinders. The number of eddies generated by the interaction depends on the ratio G/d and on the geometry of the encounter, which is given by the ratio y/g, where g = G/2. An unexpected and revealing result was the formation of a dipole vortex downstream of the two islands for values of −2 < y/g < 0, 0.25 ≤ G/d ≤ 0.4, and Re > 200, where Re = UG/ν is the Reynolds number and U is the maximum velocity of the vortex fluid in the gap. A possible mechanism is that the flow within the eddy was funneled between the two islands and, provided it had a sufficiently high velocity, a dipole formed, much like water ejected from a circular nozzle generates a dipole ring. The formation of a vortex of opposite sign to the incident vortex (i.e. anticyclonic) is in agreement with recent observations of North Brazil Current (NBC) rings interacting with the islands of St. Vincent and Barbados in the eastern Caribbean. The passage between the islands of St. Vincent and Barbados has values of G/d of approximately 0.5, hence the laboratory result
suggests that both cyclonic and anticyclonic eddies could form downstream of them.