The earliest freshwater colonisation by marine ostracods probably took place in the early Carboniferous. The first low-salinity habitats invaded by ostracods were likely ephemeral ponds near the shoreline, or fluvial, deltaic settings influenced by marine transgression events. The dominant factor in such environments is rapid change in salinity resulting from fluctuations in sea level, evaporation and precipitation, so the first freshwater colonisers are likely to have been tolerant euryhaline species. The most obvious adaptation to such environments is tolerance to low-salinity conditions. Modern studies report that ostracods possess efficient osmoregulatory mechanisms and are characterised by wide tolerance to different salinity regimes. They also exhibit other characteristics such as mixing reproductive mode, resting and/or desiccation-resistant eggs, and brood care which may facilitate colonisation of different regimes. Ostracods possess a calcified carapace containing high concentrations of calcium carbonate but shell formation may be less effective in poorly mineralised fresh waters. However, the mineral content of early Carboniferous coastal ponds was most probably sufficient to allow the development of calcified carapaces by the first freshwater species.
Introduction
Most major groups of Crustacea include both marine and non-marine species and it is apparent that the ability to colonise freshwater habitats was developed independently on numerous occasions. The first massive freshwater invasion by marine invertebrates took place in the early Carboniferous, and the most convincing records indicate that this first colonisation event occurred as a result of transgressive and regressive events (e.g. Whatley, 1990b; Tibert and Scott, 1999; Williams et al., 2006; Bennett, 2008; Bennett et al., 2012).
This is a review paper that collates and integrates the latest research results across the field of freshwater colonisation by marine ostracods in the early Carboniferous. The aim is to elucidate the sequence of events leading to colonisation of non-marine habitats, including the acquisition of physiological adaptations to new salinity regimes. In this paper I attempt to answer the question of how ostracods, organisms with relatively large shells and a high demand for calcium, were able to colonise such a low calcium environment. In addition, I seek to explain why the first massive non-marine invasion took place in the Carboniferous, when several earlier transgressive and regressive events had not resulted in colonisation of fresh waters.
Evolution
Although the valve-based record of the crustaceans supposedly belonging to the class Ostracoda dates back to the Lower Ordovician, 480 million years (Myr) ago (Salas et al., 2007), and both molecular and fossil evidences weigh in favour of even a Cambrian ostracod origin, the oldest undoubted fossil records of ostracods of the subclass Myodocopa, with their soft body parts preserved, are from the Silurian (Wenlock) deposits of England, about 430 Myr ago (Siveter et al., 2003, 2007, 2013; Williams et al., 2008). One of the oldest records of preserved soft parts in a representative of the second ostracod subclass Podocopa was described from the Lower Devonian (416 Myr) of Ukraine (Olempska et al., 2012); however, Harvey et al. (2012) described a putative ostracod mandible much earlier, from the middle to late Cambrian of western Canada. These first ostracods were marine species, living in shallow, benthic environments (Vannier and Abe, 1992; Xianguang et al., 1996; Williams et al., 2008). Earlier reports of putative ostracods (or small bivalved arthropods) from the early Palaeozoic (Bradoriida and Leperditicopida) are now regarded as equivocal and their affinities remain uncertain (Xianguang et al., 1996; Siveter, 2008; Williams et al., 2008; Vannier et al., 2001). Also the Cambrian Phosphatocopida, once classified as ostracods, according to Maas et al. (2003) are the sister group to the Eucrustacea because of their lack of specialised post-mandibular appendages.
In the middle Silurian early myodocopes started colonising the pelagic realm (Siveter et al., 1991; Vannier and Abe, 1992) from their benthic origins. By the turn of the Devonian into the Carboniferous (ca. 360 Myr ago), the first ostracods had invaded non-marine habitats (Williams et al., 2006; Bennett, 2008); however, different lineages colonised freshwater habitats at various times. Among the earliest non-marine invaders were representatives of the superfamilies Darwinuloidea (still extant) and Carbonitoidea (extinct), about 400-370 Myr ago (Vannier et al., 2003; Martens et al., 2008).
The oldest known reports of beyrichiacean palaeocopes (subclass Podocopa, order Palaeocopida) in supposedly reduced salinities are from the middle Silurian Straiton Grits Formation of southwestern Scotland, ca. 430 Myr ago (Clarkson et al., 1998; Floyd and Williams, 2003; Bennett, 2008). Unfortunately, there is a lack of the supporting evidence, such as geochemical data, necessary to confirm low-salinity conditions (Bennett, 2008).
In the early Devonian and middle Carboniferous sediments of the Brabant Massif in Belgium possible assemblages of brackish water ostracods have been reported (Bless, 1983; Bless et al., 1988). Geisina arcuata and the giant Euprimites? koeppeni (palaeocope ostracods) and Rebskeella waxweilerensis (affinity unknown), were noted in fluviatile, deltaic environments, associated with brackish-water fauna and flora (Bless et al., 1988).
The oldest deposits containing the first freshwater genus Carbonita (a representative of the podocopid ostracods with uncertain systematic affinity, see below) are recorded from the Middle Devonian Dushan Formation of Kueichow, China (Shi, 1964; Bennett, 2008). Unfortunately, this record raises many questions and requires confirmation, since the Carbonita species was listed together with common marine ostracods, and there are no details on associated microfossils and sediments, and environmental interpretation is lacking (Bennett, 2008). Moreover, it is an oversimplification and would be misleading to say that the genus Carbonita is “true freshwater”, as there is ample evidence that some species assigned to Carbonita lived in marginal marine environments (Tibert and Dewey, 2006).
A more comprehensive study which includes palaeoenvironmental data indicative of a brackish habitat comes from the early Carboniferous (Upper Mississippian) of Nova Scotia, Canada. In a range of environments from shallow near-shore marine, deltaic to low-salinity coastal ponds, brackish species of Geisina, Shemonaella, Copelandella and presumably freshwater Carbonita were found (Dewey, 1989; Tibert and Scott, 1999).
Numerous taxa of early Carboniferous ostracods from the Midland Valley of Scotland (Ballagan Formation) have been recorded. Representatives of genera such as Shemonaella, Cavellina and Paraparchites were interpreted as eurytopic species living in marginal marine and brackish settings with fluctuating salinity from 0.5 to 30. These palaeoenvironmental inferences are supported by comprehensive analysis of associated flora and fauna, sedimentology and geochemistry (Williams et al., 2005, 2006).
One of the oldest records of putative freshwater ostracods were described from the Devonian-Carboniferous boundary from the section Baihupo in Dushan of South China (Coen, 1989; Bennett, 2008). However, the identification of Gutschickia? remains uncertain, and there are too few data to confirm a freshwater or marine influenced setting. Freshwater assemblages from the middle Carboniferous (Upper Mississippian) of Virginia, including the freshwater dwelling genera Carbonita, Darwinula, Gutschickia, Pruvostina and Whipplella (for taxonomical assignments of these genera see Tibert and Dewey, 2006), were described in carbonaceous shales at the base of a marine transgression (Sohn, 1985). But again, there are insufficient data to verify freshwater conditions.
A comprehensive study of late Carboniferous (Pennsylvanian) freshwater species from the coal measures of Northumberland and Durham (Northern England) found deposits containing species of Geisina and Carbonita. The proportion of Carbonita increased through time, which reflects increasing freshwater influence (Pollard, 1966, 1969; Anderson, 1970; Bless and Pollard, 1973; Bennett, 2008). However, it is the ostracod assemblage in the Montceau Lagerstätte, from the late Carboniferous (Upper Pennsylvanian) of France (Vannier et al., 2003), that seems to be the first unequivocal freshwater setting containing Carbonita species. Ostracods were found in ephemeral freshwater ponds, with spinicaudatan arthropods (thought to be freshwater) as an associated faunal group.
In the lowest Permian deposits from Wolfcampian rocks of mid-continental North America (Kansas), four species of the genus Carbonita were reported from a freshwater pond setting (Retrum and Kaesler, 2005). The pond's source of fresh water was likely to have been from rainfall or surface flow as a result of local flooding (Retrum and Kaesler, 2005).
Ostracods belong to the genus Carbonita are a recurrent component of Carboniferous non-marine settings (Vannier et al., 2003). However, the systematic position of the first freshwater genus still remains uncertain. The carapace shape of Carbonita, and circular adductor muscle scar resemble those of some Darwinuloidea (e.g. Martens, 1998a; Martens et al., 1998; Liebau, 2005), although other authors (e.g. Anderson, 1970) referred this genus to Cypridoidea. Sohn (1985) placed Carbonita within a separate family Carbonitidae in a monotypic superfamily Carbonitoidea, which more recently was considered by Home (2004) to belong either to the suborder Darwinulocopina or to the suborder Sigilliocopina (both within the order Podocopida). Tibert and Dewey (2006) presented a taxonomic scheme of species/genera assigned to the Carbonitoidea, suggesting a polyphyly of the group. Carbonitid ostracods are considered to be a typical component of Carboniferous to Permian nonmarine assemblages across Europe and North America (Anderson, 1970; Vannier et al., 2003). Carbonita species have been recorded from numerous localities in England, Scotland, France, Germany, Canada (Nova Scotia), and USA (Vannier et al., 2003; Retrum and Kaesler, 2005). Typically carbonitids inhabited lowsalinity or non-marine coastal ponds, freshwater swamps, slow streams, or temporary pools, usually on carbonaterich mudstone (Scott and Summerson, 1943; Retrum and Kaesler, 2005). It has been suggested that they probably adopted a reproductive strategy including parthenogenesis (or heterogony) and produced desiccation-resistant resting eggs as adaptations to unstable inland water environment (Vannier et al., 2003; Retrum and Kaesler, 2005).
Background
The first non-marine environments invaded by ostracods were most probably ephemeral ponds near the shoreline, or fluvial, deltaic settings influenced by marine transgressions (e.g. Schultze et al., 1994; Tibert and Scott, 1999; Bennett et al., 2012). After the retreat of seawater, land depressions would accumulate salt water forming coastal ponds, which gradually lose active connections with the sea. As a result of rainfall, water in the ponds would become fresher. The dominant factor in such environments is rapid change in salinity resulting from fluctuations in the sea level, evaporation and precipitation, and therefore the first freshwater colonisers are likely to have been tolerant euryhaline species, because most probably water in coastal ponds became fresher quicker than organisms could acquire adaptation to new salinity regime. Many authors emphasize that physiological adaptation to abiotic conditions is crucial for successful invasion (e.g. Moyle and Light, 1996; Lee and Bell, 1999). Thus, in many cases the invader cannot survive in the new (recipient) habitat for physiological reasons, either because of their inherent biology, or because of release under the ‘wrong’ conditions (Carlton, 1996). But how can adaptation to variable salinity conditions be achieved in the relatively stable salinity of the marine environment?
Europe and the northern part of North America in the middle Devonian were equatorial, while the South Pole was located in South Africa (Walliser, 1996). In the further course of the Devonian and Carboniferous, Euramerica and western Gondwana drifted northwards and moved closer together, eventually colliding, leading to the Variscan-Hercynian orogeny (Calder, 1998; BouDagher-Fadel, 2008). This orogeny and intensive volcanism resulted in the characteristic Devonian-Carboniferous changes in the sea level and in anoxic events (Calder, 1998; Caplan and Bustin, 1999). Transgressions and regressions seem to be of importance in ostracod evolution (Whatley, 1990b), with inundation correlated with increases in species diversity and regression with gradual or rapid extinction (Braun, 1990). Moreover, Caplan and Bustin (1999) suggested that, following the upper Famennian (Upper Devonian) transgression, gradual development of anoxic benthic conditions produced highly stressful environments resulting in a high rate of extinction of benthic organisms. The extinction of many shallow-water benthic fauna in the Late Devonian left empty niches, which were filled by new taxa in the early Carboniferous. As a consequence of the Late Devonian extinctions, intensive adaptive radiation and extremely high diversification rates occurred shortly afterwards and ostracods, in addition to other groups (e.g. conodonts, ammonoids, and trilobites), rapidly increased in species diversity (Walliser, 1996).
More than fifty transgressive-regressive depositional sequences are known in Carboniferous and Permian shallow marine successions worldwide (Ross and Ross, 1988). They average about 2 Myr and range from 1.2 to 4 Myr in length. Transgressions and regressions were depositional events resulting from eustatic sea-level fluctuations that generally ranged from 100 to 200 m (Ross and Ross, 1988). As a result of early Carboniferous transgressions, shallow epicontinental seas covered large parts of Europe and Northern America (Moore, 1959; Dewey, 1989; Grossman et al., 1993). An understanding of ancient epicontinental seas is essential to our interpretation of past ecological and environmental changes (Allison and Wells, 2006). Epicontinental seas were characterised by various features: salinity was controlled by oceanic water supply, and other ionic content was influenced by the delivery of sedimentary material from terrestrial sources (Dadlez and Jaroszewski, 1994; Southard, 2007); they were typically shallow (10 to 200 m depth), but of vast extent, covering areas of up to 106 km2 (Allison and Wells, 2006). Because epicontinental seas were shallow, evaporation and precipitation brought about gradual changes in salinity. As a result of reduced tides such seas would have been more like variably salty lakes, which would have been impacted upon by lateral and vertical mixing of water (Allison and Wells, 2006). Moreover, they were sufficiently shallow that in deltaic habitats, flowing fresh water could penetrate to the bottom, thereby forming brackish conditions for benthic organisms. Carboniferous inland seas seem to provide the optimal conditions where broad salinity tolerance of benthic invertebrates could evolve.
Adaptations
The most obvious adaptation for life in non-marine environments is tolerance to low-salinity conditions. Hypersaline species must be hypoosmotic regulators, while ostracods living in fresh water are hyperosmotic regulators (Lockwood, 1962; Aladin and Potts, 1996). Marine crustaceans appear to be unspecialized in their osmotic physiology and most species tend to be characterised by a high surface permeability to both ions and water. Haemolymph concentration is almost isoosmotic with the medium, but differs in ionic composition (Lockwood, 1962). Thus, non-marine organisms, having higher ionic concentration of body fluids than the medium, tend to expend more energy regulating osmotic pressure and ionic concentration than is required for marine forms (Lee and Bell, 1999). Aladin and Potts (1996) distinguished three osmoregulation types of marine ostracods: stenohaline osmoconformers I with salinity tolerance between 30 and 36, osmoconformers II (20–40) and euryhaline osmoconformers III living in salinities from 8 to 40. They emphasised that wider salinity tolerance must be associated with the development of cell volume regulatory mechanisms (Aladin and Potts, 1996). There are several different response pathways: restricting the permeability of the body surface, active uptake or intrusion of ions, regulation of body water volume, conservation of water or salts by the excretory organs, and regulation of cellular osmotic concentration (Lockwood, 1962). Each of these adaptation pathways is probably genetically determined and develops in response to the various salinity conditions. Hypoosmotic regulation of both adult and juveniles is regulated by the excretion of salt by mitochondria-rich cells located in the non-calcified zone of the inner carapace layer (Aladin, 1983, 1984, 1987). In fresh water, hyperosmotic regulation is controlled in different ways depending on the development stage. Eggs contain salt reserves in the yolk. Juveniles absorb salt from ambient water via special cells located in the non-calcified zone of the inner shell layer. Osmoregulation of adults is supported by salt consumed with food, in addition to the reabsorption of salts in the antennal glands (Aladin and Potts, 1996).
The environment of the first freshwater colonisers was characterised not only by the wide spectrum of salinity, but also by the rapidity with which extensive changes may occur. Lockwood (1962) suggested that inhabitants of such labile conditions are often benthic, burrowing forms. Salinity fluctuations in water flowing over sand do not result in rapid changes in the interstitial water of the substratum (Reid, 1932). Most probably, burrowing forms, such as many ostracod species, could escape from exposure to the extreme changes of salinity up in the water column.
Pelagic forms are typically more efficient at dispersal than benthic species (Bilton et al., 2001; Kinlan and Gaines, 2003; Johnston et al., 2009), and most likely occurred in Carboniferous sea (Groos-Uffenorde and Schindler, 1990), but pelagic ostracods have never invaded fresh water. One possible reason is the lack of an effective strategy to cope with changing salinity conditions.
Another factor is that non-marine habitats tend to experience greater fluctuations in water temperature than marine environments. Salinity and temperature interact as stressors in their effects on tolerance—affecting metabolic rate, ion uptake, and membrane permeability. Given that temperature fluctuations are greater at higher latitudes, it has been suggested that freshwater invasion is most likely at lower latitudes (Lee and Bell, 1999). It is symptomatic that the most convincing evidence of freshwater invasion by marine ostracods derives from Western Europe and Canada—both regions that had tropical climate and were located near the equator in the early Carboniferous (Grossman et al., 2002). Other data suggest that crustaceans, generally as a group, are characterised by the ability to reduce membrane permeability (Lee and Bell, 1999). However, the permeability of ostracods to sodium ions is markedly lower than in other small crustaceans (Aladin and Potts, 1996).
The survival of an invader is also partly dictated by feeding strategy. Moyle and Light (1996) suggest that the most efficient invaders are omnivores and detritivores, because such food sources rarely seem to be limiting in aquatic environments. In lakes, also zooplanktivores seem to be successful invaders. Marine ostracods are characterized by a wide variety of feeding strategies, from suspension feeders, omnivores, detritivores, scavengers, predators, to ectoparasites (Vannier et al., 1998). However, freshwater ostracods are mainly omnivorous detritus feeders. In the Pennsylvanian coal measures of northern England Carbonita species seem to be better adapted to freshwater environments than those of Geisina. Geisina is considered to be a filter feeder, while Carbonita may have been a deposit feeder. The success of Carbonita has been inferred to be a result of its feeding strategy in the new benthic habitat (Bennett, 2008).
It has also been emphasized that asexual (parthenogenetic) reproduction facilitates invasion of new habitat, because only one egg is sufficient to colonise a new water pool (e.g. Whatley, 1990a; Martens, 1998b). But asexuals seem to have limited mechanisms of genetic and ecological plasticity, which should result in a decreasing rate of adaptive evolution (Doninck et al., 2003). Carlton (1996) suggested that genetic variability of colonist populations can be a factor determining the effectiveness of an invasion. The first freshwater invaders had to face the very changeable and dynamic environment of marginal marine or ephemeral coastal ponds. Only a tolerant form could survive. On the other hand, sexual reproduction is typically associated with the production of genetically diverse progeny (Maynard Smith, 1978), which is probably beneficial for exploitation of a new habitat type, but the success of colonisation then depends on size of “inocula”. The larger the invader population size, the greater probability of individual colonists locating each other and successfully reproducing (Carlton, 1996). Therefore, mixed reproduction, combining prevailing syngamic reproduction during unstable environmental times and parthenogenetic reproduction during stable environmental times, seems to be the most efficient for colonisation of changeable new environments (Schön et al., 2000). The vast majority of living ostracods reproduce sexually; parthenogenesis is known mostly in freshwater groups: Cyprididae, significant number of candonids, cytheroid non-marine species and all Darwinuloidea (Martens, 1998b). Dewey (1987, 1989) recorded in Carboniferous marginal marine settings of Nova Scotia Paraparchites species (Palaeocopida) that exhibited a multigenerational, progenetic parthenogenetic population. Retrum and Kaesler (2005) noticed variation in size of Carbonita specimens in different localities of early Permian deposits in Kansas, and interpreted this either as sexual dimorphism, or as a result of parthenogenetic or mixed reproduction.
It is possible that early Carboniferous ostracods had other advanced reproductive strategies that facilitated survival in marginal marine environments. Possible brood rearing of Copelandella (Beyrichiidae) was recorded from the Horton Bluff Formation of Nova Scotia (Tibert and Scott, 1999). Calcareous spheres were observed within the antero-ventral lobes (possible cruminal sacs) and were interpreted as eggs (Tibert and Scott, 1999). A similar observation for the Tournaisian (early Carboniferous) Cavellina lovatica was recorded by Gramm and Egorov (1972). The oldest fossil record concerning brood care in ostracods comes from Silurian Herefordshire deposits in England (Siveter et al., 2007). Preserved eggs and possible juveniles within the carapace of a myodocopid species were found, indicating that this reproductive strategy had developed a long time prior to the first freshwater colonisation (425 Myr ago). Brood care might ensure higher survivorship of eggs and juveniles during the high stress of passive or active transport to a freshwater setting (Tibert and Scott, 1999).
Another reproductive strategy for non-marine ostracods is producing resting and/or desiccation-resistant eggs (Carbonel et al., 1988; Bennett, 2008). Possible resting ostracod eggs were recorded from the Mesozoic (Whatley, 1990a, 1990b; Smith, 1999), but it is possible that opportunist species also acquired this strategy by the Carboniferous (Vannier et al., 2003).
The calcification problem
Ostracods are often heavily calcified and their carapace contains high concentrations of calcium carbonate. Even the nauplius, the first instar, possesses a slightly calcified duplicature of the dorsal cuticle which encases the whole body (Keyser and Walter, 2004). As in other crustaceans, the valves of the carapace are mineralised with low magnesium calcium carbonate in the form of calcite (Kesling, 1951; Sohn, 1958), and this differs from molluscan shell which consists of aragonite (Keyser and Walter, 2004).
The importance of calcium availability in controlling and limiting the post-embryonic development of ostracods has been reported in several studies (e.g. Holmes and Chivas, 2002; ). Sea water is characterised by relatively high ionic content, depending on the water temperature, with a conductivity for 35 varying from 3000 to 5500 µS cm-1 (Dera, 2003). The mineral content of fresh water is variable, but is usually lower than 1200 µS cm-1, and ionic content is influenced mainly by the mineral content of the basin's bedrock. For example, Langmuir (1997) reported that mean content of Ca and Mg in river water was 15 mg dm-3 and 4.1 mg dm-3 and for seawater -410 mg dm-3 and 1,350 mg dm-3 , respectively. Moreover, calcium in seawater occurs mainly in the form of calcium sulphate, and, in lower concentrations, as calcium carbonate. Sea water is also rich in other ions, such as magnesium in a form of magnesium chloride and magnesium sulphate (Schulz and Zabel, 2007). In fresh water dissolved ions most commonly have the form of carbonates, but generally non-marine habitats are poor in calcium and magnesium, which is reflected in the relatively thin and poorly mineralised shell of most freshwater ostracods.
Several authors (e.g. De Deckker et al., 1999; Holmes and Chivas, 2002) noticed that the Mg/Ca ratio in the ostracod carapace is strongly correlated with the Mg/Ca content of water. In addition, De Deckker et al. (1999) noticed that there is no relationship between Mg/Ca of the ostracod shell and water salinity. This suggests that for building a carapace, an appropriate calcium and magnesium content is more important than high salinity. However, this observation conflicts with the findings of other authors (e.g. Chivas et al., 1986). Uptake of Mg and Ca in ostracod valves is also positively related to water temperature; thus it was easier for the first nonmarine invaders to mineralise the carapace in the warm, tropical ponds of the early Carboniferous. Decrouy et al. (2011) noted that, in shallow waters, higher temperatures increased the Mg/Ca and DIC (dissolved inorganic carbon) concentrations of the water, which may influence ostracod shell mineralisation. Furthermore, De Decker et al. (1999) observed that the effect of temperature on the uptake of Mg is much more substantial than a change in the Mg/Ca content of the medium.
Keyser and Walter (2004) claim that the carapace of Cytheroidea (which comprises mainly marine species) is almost exclusively built of calcite. They also cited Kinser, who noted 82.7% calcium carbonate within the shell of freshwater tropical species of Chlamydotheca. Water chemistry and temperature are usually considered as being the two main factors to influence the development and survival of freshwater ostracods (Mezquita et al., 1999). Why did organisms with a strongly calcified carapace invade non-marine environments that are poor in calcium and other ions?
Why Carboniferous?
Carlton (1996) summarized that any colonisation event required the proper combination of conditions and circumstances. Invasive species are likely to continue to invade if corridors are available and conditions permit. In the Carboniferous the ostracods were not the only taxonomic group to succeed in colonizing fresh water. The massive colonisation included also bivalves, gastropods, polychaetes and other groups of crustaceans (Korejwo, 1977; Calder, 1998). The wave of transgression events provided an efficient means of transporting colonist populations. Shallow epicontinental seas covering large parts of Western Europe and North America for millions years provided the conditions that allowed the development of a wide tolerance for salinity fluctuations in changeable environments. Colonists also acquired a wide variety of other physiological adaptations in their feeding and reproductive strategies, such as brooding care and desiccation-resistant eggs, which may have led to their success in invading the new salinity regime. But why did they succeed in the Carboniferous?
It is estimated that there were at least 14 marine transgression events in the Devonian (Groos-Uffenorde and Schindler, 1990), but the most successful and permanent colonisation of fresh water took place in the early to middle Carboniferous. The probable explanation is that during the Late Devonian and early Carboniferous plants transformed the terrestrial environment. During the Carboniferous, for the first time in the history of Earth, the land was covered by tree-ferns and the first vascular plants (Calder, 1998). Because the climate was tropical and humid (Calder, 1998; Grossman et al., 2002), and the oxygen content was high (up to 35% according to Beerling et al., 1998) organic matter decay may have been rapid. High levels of organic matter delivery from the land stimulated primary productivity in inland waters (Caplan and Bustin, 1999).
In addition, the ionic content of epicontinental seas is mostly controlled by chemical composition of the bedrock of the inundated land (Dadlez and Jaroszewski, 1994; Southard, 2007). Late Devonian and early Carboniferous Western Europe and North America are characterised by sediments consisting mainly of calcareous shale, limestones and dolostones (Bless et al., 1988; Calder, 1998; Tibert and Scott, 1999). Such bedrock was most likely a considerable source of calcium and magnesium ions. Early Carboniferous shallow seas were also rich in Foraminifera: large benthic fusulinines, species of Ozawainellidae and Staffellidae (BouDagher-Fadel, 2008), their tests most likely replenished calcium compounds and other ions in bottom sediment. During marine regressions coastal ponds trapped marine waters and gradually became fresher as a result of rainfall supply. Thus, the mineral content of early Carboniferous coastal ponds might have been high enough to allow calcified carapaces to be developed by the first freshwater species. Because of the high productivity of Carboniferous fresh waters, invasion windows were open, and new faunal niches could become established.
It seems that colonising fresh water was a great success for the Darwinuloidea. During the Permian-Triassic global events most probably all marine species of this superfamily became extinct (Doninck et al., 2003), due to dysoxic transgressions, volcanism and modifications in climate and oceanic circulation (Crasquin-Soleau et al., 2004), but assemblages of continental ostracods remained unchanged between the late Permian and Early Triassic, with Darwinula as a dominant genus (Kukhtinov and Crasquin-Soleau, 1999; Crasquin and Forel, in press). Colonisation of fresh water most probably allowed the Darwinuloidea to escape extinction and to persist to this day as an exclusively non-marine group.
Acknowledgments
I am very grateful to Markus Lindholm (NIVA, Oslo), Tadeusz Namiotko (University of Gdansk) and Geoffrey Boxshall (Natural History Museum, London) for their constructive and helpful comments on the manuscript.
