Understanding the spatial integrity and connectivity of jellyfish blooms is important for ecologists and coastal stakeholders alike. Previous studies have shown that the distribution of jellyfish blooms can display a marked consistency in space and time, suggesting that such patterns cannot be attributed to passive processes alone. In the present study, we used a combination of microsatellite markers and mitochondrial cytochrome oxidase I sequences to investigate genetic structuring of the scyphozoan jellyfish Rhizostoma octopus in the Irish and Celtic Seas. The mitochondrial data indicated far higher levels of population differentiation than the microsatellites: phi(ST[MT])=0.300 vs. phi(ST[NUC])=0.013. Simulation studies indicated that the low levels of nuclear differentiation were not the result of limited power because of low levels of polymorphism. These findings, supported by palaeodistribution modelling and mismatch distribution analysis, are consistent with expansion of R.octopus from a single, limited refugium after the Last Glacial Maximum, followed by subsequent isolation, and that the discrepancy between the mitochondrial and nuclear markers is a result of the nuclear loci taking longer to reach mutation-drift equilibrium following the expansion as a result of their four-fold larger effective population size. The populations studied are probably not well connected via gene flow, and thus genetically as well as geographically distinct, although our findings also highlight the need to use a combination of organellar and nuclear markers to enable a more complete understanding of population demography and structure, particularly for species with large effective population sizes.