Marine invertebrates with long-lived planktonic larvae are assumed to have high dispersal potential and broad-scale gene flow. They are thought to disperse over long distances and to be genetically homogeneous over large areas. The periwinkle Melarhaphe neritoides which ranges along much of the European coastline, is such a species. Its long-lived planktonic larval phase lasts 4-8 weeks and previous allozyme studies have shown very little to no population genetic differentiation along the European coast, even over distances of thousands of kilometers. However, a growing number of studies based on DNA markers suggest that there is a poor correlation between pelagic larval duration (PLD) and the level of population genetic differentiation (r² = 0.29). Nevertheless, recent work reaffirms that PLD might still be a good predictor of population genetic differentiation as a function of geographic scale, if based on unbiased FST estimates and appropriate sampling sizes. We explored these conflicting views by assessing mtDNA (COI, 16S, Cytb) differentiation among five intensively-sampled populations of M. neritoides in the Azores. In contrast with previous allozyme data, our mtDNA data reveal remarkably high genetic diversity and differentiation among M. neritoides populations, even those as close as 50 km. Even at this scale, populations do not seem to share haplotypes, despite the assumed high potential of broad-scale gene flow in M. neritoides. Hence, the relationship between PLD and dispersal potential or the level of population genetic differentiation, is indeed not straightforward.