The marine environment provides many opportunities for dispersal of individuals within and among populations (Cowen and Sponaugle, 2009). Benthic invertebrate taxa usually have very limited dispersal potential as adults, whereas many species have pelagic larval stages that facilitate dispersal (Weersing and Toonen, 2009). The dispersal ability of a taxon is directed by a number of factors such as historical processes, environmental conditions, currents and life history traits such as the duration of the larval stage (Pringle and Wares, 2007; Jolly et al., 2009). These factors can lead to a variety of distribution pattens—from cosmopolitan distributions of species to the existence of populations with significant genetic differences even at close geographic distances (Hohenlohe, 2004; Hart and Marko, 2010; Derycke et al., 2013; Iacchei et al., 2013; Vergara-Chen et al., 2013).

The evolution of genetic diversity patterns depends mostly on the interplay of mutation, random genetic drift, gene flow and natural selection (Hartl and Clark, 2007). These processes can be studied through population genetics and phylogeographic approaches (Derycke et al., 2013). Over the last two decades, the development of molecular techniques has allowed the in-depth study of population structure and dynamics in benthic invertebrates (e.g., Duran et al., 2004). Mitochondrial genes and especially the cytochrome c oxidase subunit I (mtCOI) are widely used and have been proved useful for both DNA barcoding of species (http://www.barcodeoflife.org/) and population genetic analysis (e.g., Avise, 1994; Galtier et al., 2009). For instance, mtDNA data have been widely used to assess the temporal and spatial fluctuations of haplotype diversity in natural populations. Many studies investigate cryptic species along environmental gradients (e.g., Jolly et al., 2005; Barroso et al., 2010), but few focus on the population structure within a specified area (e.g., Schulze et al., 2000; Craft et al., 2010; Chust et al., 2013).

At the community level, a series of mathematical approaches and indices exist to calculate quantitative estimates of divergence. Whittaker (1960) introduced the term “beta diversity” (often referred to as turnover diversity) to describe the diversification between assemblages within a certain geographic area. Since then, 24 different (dis-)similarity coefficients have been proposed to quantify beta diversity (Legendre and Cáceres, 2013).

There is a conceptual analogy between beta diversity of community structure and genetic divergence between populations of a species. Theory predicts that environmental and stochastic processes act at all levels of the biological organization and thus cause species diversity and genetic diversity to co-vary in time and space (Etienne and Olff, 2004; Vellend and Geber, 2005). A number of studies have identified the parallel influence of local processes on species and genetic diversity (e.g., Vellend, 2003; Papadopoulou et al., 2011), as well as the mutual shaping of community diversity and the genetic diversity of its members (e.g., Booth and Grime, 2003; Whitham et al., 2003; Hughes et al., 2008). If these theories hold true, beta diversity and population genetic diversity can be expected to change at similar rates between locations. Indeed, Baselga et al. (2013) have found such correlations in communities of water beetles across Europe, and Papadopoulou et al. (2011) revealed that beta diversity and haplotype diversity of tenebrionid beetle communities on the Aegean Islands declined at similar rates with increasing distances.

Furthermore, there are strong indications that the genetic diversity of a single species can have effects on the diversity of a whole community if that species has a dominant or keystone role in the ecosystem (Treseder and Vitousek, 2001; Whitham et al., 2003). If the genetic diversity of certain species is shown to change at similar rates to the species diversity of a community, these species could be used as a proxy for time-consuming and costly whole-community analyses (e.g., Féral et al., 2003). In this paper, we analyse the genetic differences of the abundant polychaete Hermodice carunculata (Pallas, 1766) during two successive years at two locations in northern Crete (Aegean Sea). The results are combined with an analysis of ecological divergence of the total polychaete community structure between these locations. This allows us to assess to what extent the genetic diversity of a single, dominant species reflects the overall community diversity, and provides a basis for the future exploration of such species as biological indicators.

Both the mitochondrial and nuclear markers such as COI, 16S and 18S have been extensively used for the assessment of the genetic structure of populations in many geographic areas (e.g., Reeb and Avise, 1990). The dispersion potential of the species, the presence or absence of clear barriers and the effective population size are important components affecting the genetic population structure (Foll and Gaggiotti, 2006). In the case of the annelids, a rather complex pattern seems to have emerged from these factors. Schulze (2006) found that despite the absence of planktonic larvae in eunicid polychaetes, long distance dispersal is possible in at least some lineages. Within the genera of the family Amphinomidae, both ends of the spectrum have been found: Strongly divergent populations have been detected in the genus Eurythoe (considered to be cryptic species, Barroso et al., 2010), and widely distributed populations representing single species, such as those of Hermodice carunculata (Ahrens et al., 2013). Although the existing references for the reproductive biology and larval stages of amphinomids are scarce (e.g., Allen, 1957; Kudenov, 1974; Giangrande, 1997), teleplanic and rostraria larval forms are typical of the family. Both forms are planktonic and especially that of the unique rostraria with elongated feeding tentacles supports the theory of a long distance larval dispersal potential. However, Glasby (2005), based on a study of local polychaete endemism, postulated that the presence of a pelagic larval form does not necessarily imply a gene flow capable of homogenizing distant populations or preventing speciation.

The phylogenetic analysis of H. carunculata populations has revealed two main clades. These clades are separated having strong branch support values. The sequences included in Clade I reflect the long isolation of Atlantic and Mediterranean populations, having led to divergence and lineage sorting, although selective factors cannot be excluded to have added to such divergence. Clade II includes subclades consisting of individuals from both Mediterranean and from Mediterranean populations. The latter is indicative of the gene flow mechanism through which high connection rates between distant populations can be recorded. Ahrens et al. (2013) proposed that this gene flow mechanism is largely facilitated by factors such as sea water currents and ship transportation (ballast water and hull fouling), and which are capable of establishing connections between the H. carunculata populations in the Mediterranean Sea and the Atlantic Ocean.

Despite the ability of the taxon to maintain gene flow over long distances, the two Cretan populations show an important degree of differentiation according to the F_ST index, when compared to the populations within the Mediterranean and between the Mediterranean—Atlantic sites. The small geographic distance between the two sampling sites in Crete is reflected by lineage sorting, which in turn leads to fairly distinct branches for each location on the haplotype network. Among the 15 found haplotypes, only one is shared between the two sampling sites. This pattern implies a very limited or no gene flow, which could be by explained by a small effective population size (Slatkin, 1987). Effective population size is a key component affecting genetic population structure, as exemplified by differences in the case of the genetic structure between two nematode species (Caenorhabditis remanei and C. elegans) with differing effective population sizes due to different breeding behavior (Cutter et al., 2006; Charlesworth and Charlesworth, 2010). The effective population size may also be the main factor that accounts for low divergence in benthic nematodes (Bik et al., 2010) which show huge population sizes and a the high number of cryptic species, and the high divergence usually found in polychaetes with much smaller effective population sizes (Ahrens et al., 2013 and references therein).

Considering such a relatively quick lineage sorting, it is difficult to understand the occurrence of Clade II with representative samples in both the Atlantic and the Mediterranean. On the other hand, in other invertebrate species, distinct oceanographic conditions, occurring at spatial scales of just a few kilometers, can be responsible for the detected genetic differentiation in populations (Palumbi, 2004). This has also been reported for many polychaete species (e.g., Jolly et al., 2005; Schulze, 2006; Rockman, 2012). In this study, haplotype and nucleotide diversity indices values present higher values in Elounda than in Alykes. This suggests that the populations of H. carunculata in Elounda may be larger and more stable than in Alykes (Lowe et al., 2004). The less stable a population is, the higher is the genetic drift, which in turn, may cause gene variants to disappear (Scaps, 2002). This is most probably the reason for the increased number of haplotype singletons found in Alykes.

The beta diversity between the two polychaete assemblages was found to be rather high, compared to cases in which larger geographic distances or contrasted ecosystems with geographically isolated populations were analyzed (e.g., Legendre and Cáceres, 2013). Such high values of beta diversity can be considered as the consequence of habitat diversification (Whittaker, 1972). Given the apparent homogeneity of the sublittoral habitats in the region, the absence of apparent oceanographic barriers, the shallow water currents along the coasts of Crete and the relatively small distance between the two locations, they cannot be considered as geographically isolated from each other. Therefore, the past and present environmental conditions have to be considered as the major factor responsible for the rather strong divergence of these assemblages.

The two indices (F_ST and b_w) represent different components of biodiversity in a certain area. Beta diversity reflects the differences between whole communities and thus represent species diversity, whereas F_ST is a measure of population differentiation due to genetic structure. In the Cretan locations, both the differences between the two populations of H. carunculata (based on molecular diversity) and the divergence of the polychaete taxo-communities (represented by b_w) are comparatively high, given the apparent absence of any barriers. This indicates that environmental processes in the two areas must act in a similar way on both the composition of the communities and on the genetic diversity (Vellend and Geber, 2005). Such positive correlations between species and genetic diversity have been reported before (e.g., Vellend, 2003; Papadopoulou et al., 2011; Struebig et al., 2011; Baselga et al., 2013), and our results are in line with these findings. The dominant role of H. carunculata in the studied ecosystem, being an abundant carnivore, might be an additional explanation for the corresponding patterns between species and genetic divergence. Interactions between the species and the remaining community could therefore have led to a co-variation of the two levels of the biological organization (Whitham et al., 2003). If future analyses with additional locations show that these results are constant, the species could be a candidate for an indicator species, allowing rapid assessment of the divergence of two communities without the need for time-consuming identification of thousands of specimens.

Our research has highlighted a region and a taxonomic group worthy of further investigation, and is a promising approach for studies of community structuring and evolution, as well as for the determination of biological indicator species. In the present study, we have combined a standard community ecology analysis with population genetics and have successfully highlighted a scientific field of interest for follow up studies. Although the results of the present study are limited by the fact that only two locations are compared, the indices are rather constant over the two successive years, indicating that the populations and communities are not subjected to strong fluctuations over time. Furthermore, due to the high number of species in the community and the high number of individuals chosen for the genetic analysis the results are not statistically biased by a small sampling size, a common problem in similar analyses (Nazareno and Jump, 2012). Future studies should include additional locations as well as additional species and nuclear markers for the population genetics analysis. To complete this image, life history traits of the studied species as well as environmental factors such as oceanographic currents should be included to understand in detail the factors affecting the co-variation of species and genetic diversity.