The panmixia hypothesis-that all European eel (Anguilla anguilla) migrate to the Sargasso Sea for reproduction and comprise a single, randomly mating population-is widely accepted(1,2). If true, then this peculiar life history strategy would directly impact the population genetics of this species, and eels from European and north African rivers should belong to the same breeding population through the random dispersal of larvae. To date, the panmixia hypothesis has remained unchallenged: genetic studies realized on eel's mitochondrial DNA failed to detect any genetic structure(3-5); and a similar lack of structure was found using allozymes(6,7), with the exception of clinal variation imposed by selection(8,9). Here we have used highly polymorphic genetic markers that provide better resolution(10,11) to investigate genetic structure in European eel. Analysis of seven microsatellite loci among 13 samples from the north Atlantic, the Baltic Sea and the Mediterranean Sea basins reveals that there is global genetic differentiation(12). Moreover, pairwise Cavalli-Sforza and Edwards' 13 chord distances correlate significantly with coastal geographical distance. This pattern of genetic structure implies non-random mating and restricted gene flow among eels from different sampled locations, which therefore refute the hypothesis of panmixia. Consequently, the reproductive biology of European eel must be reconsidered(14,15).