Background: Factors promoting the emergence of sharp phylogeographic breaks include restricted dispersal, habitat discontinuity, physical barriers, disruptive selection, mating incompatibility, genetic surfing and secondary contact. Disentangling the role of each in any particular system can be difficult, especially when species are evenly distributed across transition zones and dispersal barriers are not evident. The estuarine seaweed Fucus ceranoides provides a good example of highly differentiated populations along its most persistent distributional range at the present rear edge of the species distribution, in NW Iberia. Intrinsic dispersal restrictions are obvious in this species, but have not prevented F. ceranoides from vastly expanding its range northwards following the last glaciation, implying that additional factors are responsible for the lack of connectivity between neighbouring southern populations. In this study we analyze 22 consecutive populations of F. ceranoides along NW Iberia to investigate the processes generating and maintaining the observed high levels of regional genetic divergence. Results: Variation at seven microsatellite loci and at mtDNA spacer sequences was concordant in revealing that Iberian F. ceranoides is composed of three divergent genetic clusters displaying nearly disjunct geographical distributions. Structure and AFC analyses detected two populations with an admixed nuclear background. Haplotypic diversity was high in the W sector and very low in the N sector. Within each genetic cluster, population structure was also pervasive, although shallower. Conclusions: The deep divergence between sectors coupled with the lack of support for a role of oceanographic barriers in defining the location of breaks suggested 1) that the parapatric genetic sectors result from the regional reassembly of formerly vicariant sub-populations, and 2) that the genetic discontinuities at secondary contact zones (and elsewhere) are maintained despite normal migration rates. We conclude that colonization and immigration, as sources of gene-flow, have very different genetic effects. Migration between established populations is effectively too low to prevent their differentiation by drift or to smooth historical differences inherited from the colonization process. F. ceranoides, but possibly low-dispersal species in general, appear to be unified to a large extent by historical, non-equilibrium processes of extinction and colonization, rather than by contemporary patterns of gene flow.