AbstractUsing 12 microsatellite loci, mitochondrial DNA sequences and previously published morphological evidence, we examined the phylogeographic and taxonomic structure of Natrix helvetica. Our results support tentatively the recognition of five subspecies: (1) N. h. helvetica (Lacepède, 1789) from Western Europe, (2) N. h. cetti Gené, 1839 from Sardinia, (3) N. h. corsa (Hecht, 1930) from Corsica, (4) N. h. lanzai Kramer, 1970 from the Po drainage and peninsular Italy, except for parts of Calabria and perhaps Apulia, and (5) N. h. sicula (Cuvier, 1829) from Sicily and parts of Calabria. The status of the subspecies from Corsica and Sardinia warrants further research. Grass snakes from the two islands are genetically deeply divergent from their continental and Sicilian conspecifics. However, the mitochondrial haplotypes of N. h. cetti and N. h. corsa are not reciprocally monophyletic, and our analyses of microsatellite data did not unambiguously support their distinctness. Natrix helvetica lanzai comprises three mitochondrial lineages and could represent more than one taxon. Grass snakes of the northern mitochondrial lineage of N. h. lanzai (mainly from the Po drainage system) were previously assigned to N. h. helvetica. We propose that the recognition of subspecies within N. helvetica reflects its genetic differentiation best. Our subspecies concept resembles the definition of Evolutionarily Significant Units (ESUs) in that subspecies should be ideally confirmed by two independent genetic lines of evidence (mtDNA, nuclear genomic markers). In contrast to species, subspecies are still fully capable of extensive gene flow and may disappear in the evolutionary process as a consequence of secondary contact. Thus, subspecies represent an early stage of (incomplete) speciation. Using the subspecies category facilitates communication within and beyond science (legislation, conservation), a clear advantage compared to the cumbersome concept of ESUs.