Numerous planktonic species have disjunct distribution patterns in the world's oceans. However, it is unclear whether these are truly unconnected by gene flow, or whether they are composed of morphologically cryptic species. The marine planktonic chaetognath Sagitta setosa Müller has a discontinuous geographic distribution over the continental shelf in the northeastern Atlantic, Mediterranean Sea, and Black Sea. Morphological variation between these populations has been described, but overlaps and is therefore unsuitable to determine the degree of isolation between populations. To test whether disjunct populations are also genetically disjunct, we sequenced a 504-bp fragment of mitochondrial DNA comprising the cytochrome oxidase II region of 86 individuals. Sequences were highly variable; each represented a different haplotype. Within S. setosa, sequence divergence ranged from 0.2 to 8.1% and strong phylogeographic structure was found, with four main groups corresponding to the northeastern Atlantic, Mediterranean Sea (including Ligurian Sea, Tyrrhenian Sea and Gulf of Gabès), Adriatic Sea, and Black Sea. Two of these (Atlantic and Black Sea) were resolved as monophyletic clades, thus gene flow between disjunct populations of S. setosa has been extremely limited and lineage sorting has taken place. The deepest divergence was between Atlantic and Mediterranean/Black Sea populations followed by a split between Mediterranean and Black Sea populations. The Mediterranean/Black Sea clade comprised three groups, with the Adriatic Sea as the most likely sister clade of the Black Sea. These data are consistent with a colonization of the Black Sea from the Mediterranean. Furthermore, a possible cryptic species was found in the Black Sea with 23.1% sequence divergence from S. setosa. Two possibilities for the evolutionary origin of this species are proposed, namely, that it represents a relict species from the ancient Paratethys, or that it represents another chaetognath species that colonized the Black Sea more recently. Even though the exact timing of disjunction of S. setosa populations remains unclear, on the basis of the geological and paleoclimatic history of the European basins and our estimates of net nucleotide divergence, we suggest that disjunct populations arose through vicariance resulting from the cyclical changes in temperature and sea levels during the Pleistocene. We conclude that these populations have remained disjunct, not because of limited dispersal ability, but because of the inability to maintain viable populations in suboptimal, geographically intermediate areas.