Delimiting species is an old systematic problem, which continues to be controversial (e.g., [1–9]). If species delimitation estimates are based solely on mtDNA, groups can be influenced by the timing of speciation and the migration rates or dispersal capabilities of the species . In addition, it is crucial to better understand species boundaries in deep-water groups. Estimation of divergence times in deep-sea faunas is a promising approach to understand events that have influenced both the evolution of these neglected marine organisms and the changes in this poorly explored environment.
Species are the fundamental units in many studies on systematics, biogeography, epidemiology, and conservation biology (e.g., [9, 11]). Species also constitute the unit for assessing biodiversity and therefore accurate identification of individuals is crucial in many areas of inquiry. In systematics and biogeographical analyses, species are frequently used as terminal taxa in phylogenetic analysis. However, relatively little effort has focused on the process of identifying and delimitating such species. In addition, there are no universal criteria by which species should be delineated and identified, and both, non-tree based (measures of gene flow) and tree-based approaches have been applied [4, 6, 7, 12–15].
In octocorals, species identification has traditionally been based on external morphology (axis, branching pattern, calyx morphology, polyp arrangement, surface texture, coloration, etc.) and sclerite composition [16–20]. The taxonomy of precious corals is particularly enigmatic, because many species were described based on small deep-sea fragmentary samples, limiting our understanding of their real intra-colony and intraspecific variation . In addition, some descriptions are imprecise and the whereabouts of the type material remains unknown.
Molecular species delimitation usually uses tree-based methods to assess monophyly  or more recently have applied genealogical-based coalescent methods (e.g., [2–5, 9, 11, 22–24]). Tree-based methods reconstruct the evolutionary relationships among individuals and look for reciprocal monophyly, but are often dependent on the method/model of tree inference and it is not uncommon that single-gene trees differ from concatenated gene-based inferences. Thus, concatenation of markers and multilocus coalescent approaches have been favoured more recently (e.g., [23, 25]). Species-tree inference and estimation of divergence times from a single locus (often a mitochondrial gene) and multi-individual data are possible in a species tree framework , although these analyses are usually carried out using multilocus datasets . However, coalescent-based approaches can show signal for lineage divergence despite the lack of monophyly in gene-trees, which can be common in recently diverged species . Time-calibrated phylogenies have also the advantage of using general speciation models, which examine species boundaries from sequences by identifying evolving lineages bridging the coalescent with speciation in the phylogeny .
The particular interest on Coralliidae phylogenetics responds also to the intricate evolution of its axial characters and the questionable validity of its two currently recognized genera, in addition to aspects of their biology and conservation, as most species are of commercial interest for the jewelry industry. The diversity of colonial forms among the various species of precious corals corresponds with various shapes of the supporting axis. Branching varies from sparse, in Corallium abyssale Bayer, to profuse, in Paracorallium japonicum Kishinouye. Terminal branchlets vary from stout and blunt to slender and delicate. The surface of the axis may be smooth or distinctly marked by narrow longitudinal grooves underlying the coenenchymal canals and some species include pits [28, 29]. Coenenchymal sclerites are basically of the radiate type, slender rods and smooth or rough double clubs [30–32].
Corallium and Paracorallium, the two accepted Coralliidae genera, comprise 17 and 7 described species, respectively. Most have been documented from the Pacific Ocean and only two species are from the eastern North Atlantic [21, 29, 31]. Coralliidae is closely related to Paragorgiidae [33–35], whose members lack a supporting axis. The most complete molecular phylogenetic hypothesis for Octocorallia  shows that Scleraxonia, the suborder containing Coralliidae, is polyphyletic. To date, limited molecular systematic studies of Coralliidae [33, 35] have provided evidence for major discrepancies between phylogeny and generic taxonomy. Additional studies of precious corals include genetic surveys, radial growth rates, dating, and recruitment estimates [37, 38], but little is known about their phylogenetic relationships. Furthermore, all precious corals are currently threatened not only by the jewelry industry but also the trawling activity on seamounts and other deep-sea rocky habitats [39–42]. However, not all precious coral species have been included in the CITES list (the Convention on International Trade in Endangered Species of Wild Fauna and Flora) due to difficult identification and uncertainty about population status.
In this study, before understanding the phylogenetic relationships within Coralliidae, it was important to delineate species boundaries with particular interest in the precise membership of the type species of the family, Corallium rubrum Linnaeus, and to re-evaluate the validity of the genus Paracorallium Bayer & Cairns. Here, we provide an extended phylogeny of Coralliidae based on a large geographical and broad taxon sampling and using data from five mitochondrial regions. We use these data for a genealogical approach to assess the significance of clustering and lineage divergence and delimit species of precious corals.