Panmictic species pose particular problems for conservation because their welfare can be addressed effectively only on a global scale. We recently documented by means of microsatellite analysis that the European eel (Anguilla anguilla) is not panmictic but instead shows genetic isolation by distance. In this study, we extended the analysis to the American eel (A. rostrata) by applying identical analytical procedures and statistical power. Results obtained for the American eel were in sharp contrast with those obtained for the European eel: the null hypothesis of panmixia could not be rejected, and no isolation by distance was detected. This implies that the species must be managed as a single population. Using Bayesian statistics, we also found that the effective population sizes for both species were surprisingly low and that the populations had undergone severe contractions, most probably during the Wisconsinan glaciation. The apparent sensitivity of eels to climatic changes affecting the strength and position of the Gulf Stream 20 000 years ago is particularly worrying, given the effects of the ongoing global warming on the North Atlantic climate. Moreover, additional short-term stresses such as surging glass eel prizes, overfishing and lethal parasitic infections negatively affect eel population size. The fascinating transatlantic migration and life cycle of Atlantic eels is also their Achilles' heel as these negative short- and long-term effects will probably culminate in a fatal synergy if drastic conservation measures are not implemented to protect these international biological resources.