Small invertebrates have a large potential for analysis of spatial patterns of diversity, providing ideal candidates to test for models in biogeography and macroecology (Witman and Roy 2009, Tuya et al. 2010). However, understanding the diversity of such organisms is hampered by lack of comprehensive data on species distribution and by a difficult, and unreliable, taxonomy (Larsen 2005). A partial solution to this problem is provided by analyses of mtDNA, which have previously proven useful in obtaining reliable data (Larsen and Froufe 2013).

Pleistocene glaciations are known to have played a major role in shaping the phylogeographic patterns of many organisms, particularly during the last glacial maxima (Väinölä 2003, Hewitt 2004, Maggs et al. 2008). Recently, gene flow between the Pacific and the Atlantic has been detected for several marine taxa; it is generally accepted that dispersal often goes in the Pacific-Atlantic direction (Teeter 1973, Meehan et al. 1989, Palumbi and Wilson 1990, Palumbi and Kessing 1991, Taylor and Dodson 1994, Rawson and Hilbish 1995, van Oppen et al. 1995, Väinölä 2003). Gene flow related to inter-glacial periods is suggested for a number of invertebrates, including littoral bivalves (Väinölä 2003), sea urchins (Addison and Hart 2005), crabs (Harrison and Crespi 1999) and even peracarids, such as mysids (Audzijonytė et al. 2005). These findings show that the entire northeast Atlantic region has played an important role in the diversification of the marine biota, especially in the case of cold-water species. Some authors, however, suggest that the situation is a bit more complicated, including dispersal and gene flow going in both directions (Hardy et al. 2011).

There is current consensus that species from temperate regions had to contract their distribution ranges into warmer southern areas during glacial periods, whereas during inter-glacial eras, organisms were able to recolonize warming northern areas (Hewitt 1996). Such changes have resulted in a latitudinal gradient of genetic divergence, with higher levels of diversity at southern locations, which might have functioned as glacial refugia (Luttikhuizen et al. 2008, Kettle et al. 2010, Xavier et al. 2012). For example, in the eastern Atlantic, the Macaronesian archipelagos, the Iberian Peninsula, and the Atlantic coasts of North Africa have been suggested as such glacial refugia areas (Coyer et al. 2004, Maggs et al. 2008, Kettle et al. 2010).

Tanaidaceans are among the most ecologically important and diverse benthic micro-crustaceans in the marine realm. The order currently contains more than 1000 species; the discovery of new species has been prominent in the last decade (Larsen 2005, Blazewicz-Paszkowycz et al. 2012, Bamber and Błazewicz-Paszkowycz 2013). Yet, phylogenetic affinities of tanaidaceans are still poorly understood (Larsen and Wilson 2002, Bird and Larsen 2009). Tanaidaceans can occur at high densities, often exceeding 10000 individuals per m², with some reports of over 140000 individuals per m² (Delille et al. 1985). Despite the obvious ecological importance displayed by such population densities, tanaidaceans are neglected in most ecological surveys, though several ecologists have acknowledged their relevance (Sokolova 1959, Pires 1980, Tuya et al. 2010), for example to monitor environmental impacts (Riera et al. 2011). Species of the genus Zeuxo (Tanaidacea: Peracarida) are brood-caring, non-swimming, and semi-sedentary animals for which long-distance dispersal is thought to be possible only by rafting, human transport via ballast water, or fouling (Larsen 2005). Therefore, their low dispersal potential makes them a good model for studying past vicariance events associated with glacial survival.

The identification of cryptic, or sibling, species is a well-known problem in systematics (Knowlton 1993). Traditional classification has focused mainly on morphological comparisons of organisms, sometimes resulting in conflicting results when the homology of morphological characters was difficult to establish. For example, the morphological differentiation among species of tanaidaceans is notoriously difficult, because of factors such as sexual and ontogenetic polymorphism, as well as the presence of cryptic (and sometimes even sympatric) species (Larsen 2001). This is quite apparent for tanaidaceans, since even family-level diagnostic characters might change with development and gender (Larsen and Wilson 1998, Larsen 2001). This paper aims to address these problems of species identification of cryptic species complexes based on morphological characters by combining these with molecular information.

The general purpose of this study was to investigate the phylogeographic patterns of certain species of Zeuxo, and in particular what has been thought to be a single species, Z. normani (Richardson 1905), which has an almost cosmopolitan distribution. This species is recorded from all over the world, including Australia (Edgar 2008), Bermuda (Greve 1974), California (Richardson 1905, Sieg 1980), Japan (Shiino 1951), Sri Lanka (Stebbing 1905) and Spain (Cacabelos et al. 2010). Such a cosmopolitan distribution is highly unlikely for organisms with low dispersal potential, such as tanaids (Larsen 2005), so the presence of a species complex is expected. To compare genetic divergence between species of the genus from different parts of the world, we include analysis of Z. holdichi (Bamber, 1990), Z. exsargasso (Sieg, 1980) and two additional new species from Korea and Turkey. In particular, we compare diversity at the species level by coupling traditional taxonomy and DNA analyses. To accomplish this objective, we examined DNA sequence variation using three different genetic markers, one mitochondrial (COI), one nuclear large subunit ribosomal DNA sequence (28S rRNA), and one nuclear marker (H3).

This study describes, for the first time, the genetic distribution pattern of species within the order Tanaidacea, which have low dispersal potential. Tanaidacean species can thus be prone to maintaining signatures of genetic drift caused by vicariant events, as seen for other organisms (Petit et al. 2003, Pelc et al. 2009).

The genetic analysis indicates that a ‘Zeuxo normani’ species complex exists (Fig. 1), spanning the Pacific and the North Atlantic. It furthermore indicates that there is a closer genetic relationship between the Pacific species and Z. holdichi than between Z. holdichi and other North Atlantic species. The Zeuxo normani species complex thus includes Z. normani California (Sieg 1980), Z. cf. normani, Japan (Drumm 2010), Z. cf. normani, Australia (Edgar 2008), Z. sp. A, Korea (this study), and Z. holdichi, Spain and France (this study).

The long-distance dispersal found in certain species with otherwise low dispersal abilities, such as Z. exsargasso, is thought to depend on rafting on floating substrata, particularly on algae (e.g. the genus Sargassum) (Sieg 1980). Thus, the general absence of genetic divergence throughout the distribution of Z. exsargasso (albeit not examined over the entire distribution of the species) could suggest that randomness and regional hydrographic patterns can promote high population connectivity via the Gulf Stream. It has been suggested (Thiel and Haye 2006) that there may be convergence zones where multiple rafts arrive, thus favouring high levels of gene flow between distant populations and erasing the genetic signatures of founder effects. Such ‘arrival’ zones for eastward dispersal are likely to be the Macaronesian archipelagos (Bamber 2012a).

Zeuxo holdichi (part of the ‘Z. normani’ species complex) may have retreated to refugia in the south (in northern Spain, Brittany and the English Channel) during the last glaciations, as was proposed for many cold-water taxa (e.g. Kettle et al. 2010). However, this species does not seem to be able to recolonize northern latitudes. Furthermore, it appears that a discontinuous interbreeding between the North Atlantic species (Z. holdichi) and the Pacific ‘parent’ species (Z. sp. A from Korea and Z. cf. normani from Japan) allow for morphological observable separation of this species (Z. holdichi): this is plausible if we accept that gene flow is in the Pacific-Atlantic direction. In contrast, while the genetic COI divergence (see Table 2) between Z. cf. normani from Japan and Z. sp. A from Korea is on a similar level (10.7%) to that of Z. holdichi (11.2% and 12.5%, respectively) only a marginal morphological difference can be found between these species according to the description by Shiino (1951). Importantly, this shows that regardless of small genetic (COI) divergence, morphological difference can either be present or absent.

The molecular analyses showed, as expected, the same three highly supported clades within Zeuxo spp., although the relationships between them lacked strong support. It also showed that different species of Zeuxo appear to have a genetic divergence in the COI gene, ranging from 10.7% between Z. sp. A (Korea) and Z. cf. normani (Japan) to 28.2% to 28.5% between Z. sp. B (Turkey) and Zeuxo exsargasso (see Table 2). Although few genetic studies have been made on tanaidaceans, a 20% divergence has been reported between several species of Leptochelia (Drumm 2010, Larsen et al. 2012, Larsen and Froufe 2013).

The closer genetic similarity of Z. holdichi from the Atlantic coast of Spain to the Pacific Z. sp. A from Korea (11.2%) and Z. cf. normani from Japan (12.5%), compared with that of the geographically much closer Atlantic species, Z. exsargasso (28.2%-28.5%), indicates the presence of a Z. normani species complex spanning the Pacific and North Atlantic. This close genetic similarity of Z. holdichi to the North Pacific members of the Z. normani complex suggests a previous inter-glacial colonization. We then speculate that the populations of Z. holdichi from the Atlantic coast of Spain and France are remnants of such a circumpolar distribution.

Our results point out some issues on the current diagnostic characters for species of Zeuxo that were previously found to be unstable. First, it is well known that the number of uropodal articles (Fig. 2) can vary within species of Tanaoidea though it has still been accepted as one of the most important characters for species identification since the monograph by Sieg (1980). However, already Kudinova-Pasternak (1989) mentioned variations of this character in Z. beringi, and both Bamber (1990) and Edgar (2008) confirmed this observation in other species of Zeuxo. Similarly, this study, supported by genetic confirmation, proves that this character can be variable within a species and, indeed, even within an individual. A very thorough study on this character in the closely related genus Zeuxoides (Sieg, 1980) has been conducted by Bird (2008), demonstrating a clear correlation between body size (instars) and uropod article number. Our study has shown that variation can even exist between the left and right uropod of the same specimen (Fig. 2C), regardless of the body size. Therefore, the value of the number of uropodal articles as a character for species identification should be applied with caution. We have seen similar cases of uropod article variation within individuals of other tanaidaceans (Larsen et al. 2012, Larsen and Froufe 2013), and this is thus too often to just be considered a rare aberration. Second, the Parazeuxo subgenus diagnosis (Sieg 1980) depends mainly on the pleopod outer setation character and the number of spiniform setae on the pereopodal merus and carpus. This study has shown that the pleopod endopod setal character (Fig. 3) is variable in Z. exsargasso and Z. holdichi, thus making the subgenus diagnosis invalid. In both of these species, the spiniform armament of the pereopod carpus also seems to vary, and should be applied, again, with caution. Third, in recent years, some authors (Edgar 2008, Bamber 2012a, Kakui pers. comm.) have suggested that pigmentation could be used for species identification without the need for dissection. However, our results do not completely support this premise (Figs 4 and 5), at least for Z. exsargasso and Z. holdichi. Since both the morphological and genetic analysis confirmed the conspecificity of these respective specimens, this suggests that some species have variable carapace pigmentation patterns. No specific patterns were recorded relative to gender, while younger developmental stages appear to have the pigmentation concentrated on the anterior part of the carapace.

In summary, the conclusions of this study are the following: 1) The assembly of the northeast Atlantic comprises species that originated from both the North Pacific and the northwest Atlantic; 2) traditional species identification clearly underestimates diversity; and 3) what have been perceived as reliable diagnostic morphological characters can often be variable.

Hence, more detailed morphological, distributional and molecular surveys, with an emphasis on the intertidal assemblage, are needed to answer the questions of wide distributions records of non-swimming species without a dispersal phase. Clearly there are at least two proven possibilities: frequent rafting events (Larsen 2005, Bamber 2012a); and the presences of multiple, morphologically almost inseparable species known as species complexes (Larsen 2001, Bamber 2010, 2012b).