Distribution of European and African species of genus Diaptomus ( Copepoda : Calanoida : Diaptomidae ) : a track analysis

The distributional patterns of 13 species of Diaptomus Westwood, 1836 were analyzed using the panbiogeographical method of track analysis. Locality records were compiled from the literature and mapped for the construction of individual tracks for each species. These tracks were superimposed to find the generalized tracks. Four generalized tracks were found: (1) Siberia, Central Europe, and Iceland; (2) Northern Italy, southern France, central Spain, northern Algeria, and northern Morocco; (3) Southern France, central Spain, and northern Morocco; (4) Southern Italy, Sicily, and Albania. Five biogeographic nodes were found: (A) Southwestern Iberia, (B) Southeastern Iberia, (C) Central Iberia, (D) Cantabria, at the intersections of generalized tracks 2 and 3; and (E) Italian Peninsula Islands, at the intersection of generalized tracks 2 and 4. The main massing of the species of Diaptomus studied is located in the Iberian peninsula, where six of the species do occur. A model based on the fragmentation and differentiation of already widespread ancestors during the late Mesozoic and early Cenozoic, related to the opening the North Atlantic Ocean and the formation of the Mediterranean Sea, is proposed as a most parsimonious explanation for the observed patterns of geographic vicariance.


Biogeography of Diaptomus in Europe and Africa
Nauplius, 26: e2018028 inTroduCTion The genus Diaptomus Westwood, 1836 (Calanoida) comprises about 80 species of copepods characterized by the presence of a single eye spot and very elongated first antennae, exceeding body length.They occur in rivers and large freshwater lakes throughout the northern hemisphere.
The fossil record of Copepoda dates back to the late Carboniferous (Selden et al., 2010), but although there are already order-level phylogenies that corroborate the monophyletic status of Calanoida (Blanco-Bercial et al., 2011;Eyun, 2017;Khodami et al., 2017), no complete phylogeny of the Diaptomidae has been published to date.Thum (2004) provided a molecular phylogeny based on 18S rDNA for selected North American genera of Diaptomidae.Mookkaiah and Ravichandran (2016) and Sivakumar et al. (2016) presented molecular phylogenies of several diatoptomid species, including some of those mentioned in the present study.Albeit still limited, these phylogenetic hypotheses offer a useful framework for the analysis of the evolution of calanoid copepods in relation to their distribution over geographic space.
Because of their small size, mode of reproduction, and dormant stages resistant to desiccation, copepod species have been believed to possess cosmopolitan distributions as a consequence of high rates of passive dispersal by winds and animal vectors, especially waterfowl and migratory birds (Maguire, 1963).This has led to speculation about 'colonization waves' (Boxshall and Jaume, 2000), 'glacial refugia' (Marrone et al., 2017), and other dispersalist narratives to explain the biogeography of copepods.However, empirical evidence has not supported the cosmopolitanism of copepods, suggesting instead that these crustaceans have in fact limited dispersal rates and display geographical distribution patterns with considerable degrees of endemism ( Jenkins and Underwood, 1998;Bohonak and Jenkins, 2003;Boxshall and Defaye, 2008;Marrone et al., 2013).In the light of these findings, most recent studies of the geographical distribution of copepods have adopted a vicariant approach, under the paradigm of panbiogeography.
As first developed by Croizat (1958;1964), and later expanded and quantified by New Zealand researchers (Page, 1987;Craw, 1989;Henderson, 1989;Craw et al., 1999;Heads, 2012).Panbiogeography has been recognized as one of the main research programs in historical biogeography (Morrone and Crisci, 1995;Crisci, 2001;Crisci et al., 2003), including the biogeography of freshwater organisms (Bănărescu, 1990).The panbiogeographic method of track analysis consists of connecting the mapped locality records of different each taxa by means of lines of minimum distance, which defines the individual tracks, corresponding to the sector of geographical space where each taxon has evolved.When the individual tracks overlap for several groups, a generalized track is defined, suggesting a common history for the entire biota (Craw et al., 1999).
Track analysis of distributional patterns have been performed for selected groups of copepods: Jamieson (1998) analyzed the distribution of four species of Boeckella Guerne andRichard, 1889 in New Zealand, Menu-Marque et al. (2000) studied the distribution of this same genus in South America, and Mercado-Salas et al. (2012) studied the distribution of the American species of Eucyclops Claus, 1893, using panbiogeographic methods.However, so far the panbiogeographic method has not been applied to analyze the distribution patterns of the widespread genus Diaptomus.Marrone et al. (2017) studied the distribution of Western Palearctic Diaptomidae using a macroecological approach which took into account current and historical (paleoclimatic) factors.
In this paper, the geographic distributions of 13 species of Diaptomus occurring in Europe and northern Africa were mapped and analyzed using the panbiogeographic method of track analysis, with the aim of finding common distribution patterns and attempting to correlate these patterns with vicariant events related to the tectonic history of the region.

MaTerials and MeThods
Geographic distribution data for the species included in this study were compiled from the relevant literature and stored in an electronic spreadsheet in angular degrees format.A total of 395 occurrence records were obtained for all 13 European and African species of Diaptomus (Tab.1).

Biogeography of Diaptomus in Europe and Africa
Nauplius, 26: e2018028 (Heads, 2004) were determined at the intersection of two or more generalized tracks.The main massings (geographic concentrations of diversity) were assessed by counting the numbers of species in each cell of a 1 x 1 degree grid using DIVA-GIS.

resulTs and disCussion
The species studied and the number of occurrence records obtained for each are listed in Table 1.Distribution maps and individual tracks for each species are presented in Figures 1-10.
Five biogeographic nodes were determined: (A) Southwestern Iberia, (B) Southeastern Iberia, (C) Central Iberia, (D) Cantabria, at the intersections of generalized tracks 2 and 3; and (E) Italian Peninsula Islands, at the intersection of generalized tracks 2 and 4 (Fig. 11).
Diaptomus castor is the most widespread species, occurring from Iceland to northwestern Africa, followed by D. cyaneus which occurs from Cantabria to northwest Africa.The main massing of the species of Diaptomus included in the present study (Fig. 12) is located in Iberia, where six of the species occur.
The scattered highly localized endemics of northern Eurasia (D. barabinensis, D. charini, D. kostromanus, D. zografi) stands in contrast to the wider range of other species.The distribution of D. falsomirus (Fig. 5) appears to be centered on the Black Sea basin and is also allopatric to all other species (which are all centered west or north).Likewise, the distribution of D. mirus Lilljeborg in Guerne and Richard, 1889 is disjunct in north and central Eurasia (Fig. 8).
The distribution patterns of the species of Diaptomus included in the present study, as revealed by track analysis, suggest that two major geotectonic events had a role in shaping the evolution of these species, namely the formation of the Mediterranean Sea and the formation of Iceland.
The Mediterranean Sea has a very complex geological history, comprising not only the formation of the basin by the convergence of the African and

Biogeography of Diaptomus in Europe and Africa
Nauplius, 26: e2018028 Eurasian plates during the Late Triassic and Early Jurassic, but also several cycles of partial or complete desiccation during the Messinian age of the Late Miocene (Garcia-Castellanos and Villaseñor, 2011).This may explain the biogeographic nodes centered in the southern Iberian peninsula and around the Strait of Gibraltar, a region that marks the sector of geographic space where biotas have fragmented and coalesced during such cycles.These patterns are also corroborated by the molecular phylogenetic tree of Mookkaiah and Ravichandran (2016), where D. mirus, D. cyaneus and D. kenitraensis, D. castor form separate clades which are consistent with the observed patterns of geographic vicariance.These branching sequences are interpreted here not as separate dispersal events but as the 'sequence of differentiation in an already widespread ancestor' (Heads, 2009).
Generalized track 1, formed by the distributions of D. castor, D. glacialis, and D. rostripes is component of a standard track first identified by Croizat (1958) which includes many more elements of the boreal biota.These three species are recorded from Iceland (with just one record for D. castor in the southwestern part of the island).
Iceland lies on the divergent boundary between the North American and Eurasian tectonic plates, as well as above a hotspot, the so-called Iceland plume, which is believed to have formed the island itself (Mjelde et al., 2008;Torsvik et al., 2015).Iceland is a relatively young island, first appearing over the ocean surface about 16 Myr ago (Foulger, 2006), and this might suggest long-distance dispersal as an explanation for the presence of these three species of Diaptomus on Iceland.However, an alternative explanation which does not rule out vicariance is possible.First, the plume model is debatable (Heads, 2009): it may be that Iceland is not even underlain by a lower mantle plume, and volcanic activity in the island could result from processes restricted to the upper mantle and related to plate tectonics (Foulger and Anderson, 2005;Foulger and Natland, 2003;Foulger 2010;Heads, 2009).Second, island age cannot be taken as an absolute criterion for dating the age of taxa (Heads, 2009), as old taxa can survive as metapopulations on ephemeral islands of younger age formed at plate margins and fissures and now submerged (for example, on the Faroe-Rockall Plateau).

Biogeography of Diaptomus in Europe and Africa
Nauplius, 26: e2018028 The biogeographic nodes cluster to the central and western Mediterranean.Nodes A and B lie on the edge of the western Mediterranean and may be related to the reestablishment of the connection between this basin and the Atlantic Ocean through the Strait of Gibraltar by the Zanclean flood around 5.3 Myr ago (Garcia-Castellanos et al., 2009).One node (D) is at the Pyrenees so that might suggest that the tectonic compression of the Pyrenees and its age might be a factor in differentiation at this node.The Pyrenean chain achieved its present configuration due to the collision between the microcontinent Iberia and the southwestern part of the European Plate (i.e., Southern France), that approached in the onset of the Upper Cretaceous (Albian/Cenomanian), about 100 Myr ago, and collided during the Paleogene (Eocene/ Oligocene), around 55 to 25 Myr ago (Choukroune, 1992).The region is bounded by major faults, and such orogenic zones are known to be associated with major biological disjunctions (Croizat, 1958;1964;Heads, 1989).The node in Italy (E) seems to represent a boundary for distributions further west and likewise the tectonic activity in the region is pertinent to explain biological disjunction patterns in this region.This node is associated with the Apulian/Adriatic Plate, a tectonic microplate that separated from the African Plate during the Mesozoic, and generalized track 4 coincides with the Calabrian arc which marks the microplate boundary (Devoti et al., 2002).
Recently, Marrone et al. (2017) presented an analysis of the distribution patterns of diaptomid copepods in the Palearctic, which these authors explained on the basis of effects of recent (i.e., of Holocene age) climatic changes and post-glacial 'recolonizations' from putative 'refugia' in Western Europe.But this interpretation is not supported here, and instead the observed distribution patterns are explained in the light of much older geotectonic events and involved no long-distance dispersal over putative 'barriers' .
Vicariance promoted by geotectonic events extending back to the Mesozoic has played a most important role in shaping the distribution of the species of Diaptomus included in this study.This process give rise to analyzable patterns affecting whole biotas (the generalized tracks) and best explains the biogeographic connections observed between the copepod fauna of continental Europe and Iceland, as well as of the Iberian and Italian peninsulas.A model based on the fragmentation and differentiation of already widespread ancestors during the late Mesozoic and early Cenozoic, related to the opening the North Atlantic Ocean and the formation of the Mediterranean Sea offers a most parsimonious explanation for these patterns.

aCKnoWledGeMenTs
We thank John Grehan and Michael Heads for helpful comments and useful suggestions that contributed to improvements in the manuscript.

Figure 12 .
Figure 12.Centers of diversity (main massings) of the species of Diaptomus included in this study.Yellow: three species, light green: two species, dark green: one species.

Table 1 .
Species of Diaptomus included in this study and number of occurrence records and respectives references for each one.