Abstract
The anadromous–freshwater three‐spined stickleback ( Gasterosteus aculeatus ) system allows for inferring the role of adaptation in speciation with a high level of accuracy because the freshwater ecotype has evolved multiple times from a uniform anadromous ancestor. A cause for concern is that independent evolution among drainages is not guaranteed in areas with a poorly resolved glacial history. This is the case for the west European great rivers, whose downstream valleys flanked the southern limit of the late Pleistocene ice sheet. We tested for independent and postglacial colonization of these valleys hypothesizing that the relationships among anadromous and freshwater sticklebacks correspond to a raceme structure. We compared the reduction in plate number accompanying this colonization to the genetic differentiation using 13 allozyme and five microsatellite loci in 350 individuals. Overall microsatellite differentiation ( F ST = 0.147) was twice as large as allozyme differentiation ( F ST = 0.066). Although habitat‐specific gene flow may mask the ancestral relationships among both ecotypes, levels of microsatellite differentiation supported the hypothesis of raceme‐like divergence, reflecting independent colonizations rather than the presence of two distinct evolutionary clades. Under an infinite alleles model and in the absence of gene flow, the observed freshwater divergence might be reached after 440 (microsatellites) to 4500 (allozymes) generations. Hence, the anadromous–freshwater stickleback system most likely diverged postglacially. We conclude that the reduction in plate number in two freshwater basins probably occurred independently, and that its considerable variation among populations is not in agreement with the time since divergence.