The paleomagnetic analysis of samples from three Late Cretaceous through Eocene sites in southwesternmost Nicrague yields pole positions close to the present pole. The results from five equivalent sites in the Pacific coastal zone of Costa Rica, on the other hand, suggest a 14¿ northward translation in post-Eocene time. This movement cannot have been initiated by the development of the Cocas-Nazca spreading center and it seems likely that the extensive Late Miocene plutonism in Costa Rica defines the time of collision of the Nicoya Complex (sensu largo) and the Costa Rican island arc. The Nicaraguan as well as the Costa Rican data are very different from paleomagnetic results from Hunduras, and the tectonic boundary between northern and southern Central America, therefore, must lie to the north and east of the Nicaraguan sites.