In Australia's arid and semiarid zone, which covers about two-thirds of the continent (Martin 2006), permanent lakes are scarce and even river systems can be reduced to a series of permanent water holes for extended periods of time (Williams 1981; Faulks et al. 2010b). In these areas, temporary water bodies are important freshwater habitats. Various types of temporary water body can be distinguished, including turbid claypans, swamps with varying vegetation (e.g., black box or lignum swamps), large temporary lakes, and artificial farm dams (Williams 1981; Kingsford and Porter 1999). After rainfall, these pools usually fill from local runoff, but they may also be filled by flood or riverine water during severe flooding events and join up to form creeks or rivers for short periods of time (Timms 1999; Timms and Sanders 2002). The density of such pools across the landscape can be very high, as is the case in the catchment areas of the Paroo and Warrego Rivers in East Australia on the border between Queensland and New South Wales (part of the Murray–Darling Basin), where up to 26% of the land surface can be covered by claypans alone (Timms 1999). As a result, these areas are also important feeding and breeding grounds for water birds (Timms 1997; Kingsford and Porter 1999).
In addition to organisms, which also occur in other freshwater habitats, the fauna of temporary water bodies comprises a range of characteristic organisms (Timms 1999). A number of recent studies have investigated the phylogeography and genetic structure of aquatic organisms in Australia (reviewed in Hughes et al. 2009), but they all focused solely on taxa (fish, mollusks, and crustaceans) living in streams (Carini and Hughes 2004; Huey et al. 2006; Faulks et al. 2010a,b), permanent water holes (i.e., remnants of rivers during drought; Hughes et al. 2004; Carini and Hughes 2006), or permanent groundwater springs (Murphy et al. 2010). Strong genetic structuring with little to no gene flow was revealed in many species inhabiting groundwater springs or water holes (Hughes et al. 2004; Carini and Hughes 2006; Murphy et al. 2010), and riverine and some water hole species were usually structured genetically according to the main drainage systems (e.g., the Murray–Darling Basin, the Lake Eyre Basin, or the Bulloo), with no indication of ongoing gene flow across drainage system borders (Carini and Hughes 2004; Hughes et al. 2004; Huey et al. 2006; Faulks et al. 2010a,b). This is not surprising as these aquatic species can only disperse along interconnected waterways. However, all the studies in question neglected organisms that live exclusively in temporary water bodies, despite them being important components of the freshwater fauna of the arid zone of Australia (Kingsford and Porter 1999). Many of these organisms have resting or dormant stages, which survive the long drought periods in the sediment of dried-up pools. Among the most typical and common species in temporary pools are the “large branchiopod” crustaceans, namely tadpole shrimps (Notostraca), fairy shrimps (Anostraca), and clam shrimps (Spinicaudata and Laevicaudata), all of which produce “resting eggs” (actually cysts as these eggs are already fertilized) that need a desiccation period before hatching can be induced in a subsequent wet period (Brendonck 1996). Not only do resting eggs reestablish the active population within each pool, they are also the most important means of dispersal. Dispersal of resting eggs is always passive and may be mediated by wind, floods, or biotic vectors such as waterfowl (Bilton et al. 2001). In this latter case, the eggs either stick to the birds’ feathers or are carried internally after the ingestion of egg-bearing females. This mode of dispersal is not available to other aquatic taxa such as fish or mollusks and allows “large branchiopods” to disperse among unconnected water bodies. As aquatic species with resting eggs are generally among the first colonizers of new freshwater habitats, it has been assumed that the dispersal potential of resting eggs is high (Bilton et al. 2001). Most genetic studies, however, have revealed pronounced genetic differentiation even between populations inhabiting pools only a few kilometers apart (Vanoverbeke and De Meester 1997; Brendonck et al. 2000; Hulsmans et al. 2007; Thielsch et al. 2009; reviewed in De Meester et al. 2002). This dispersal-gene flow paradox has largely been ascribed to persistent founder effects (Boileau et al. 1992) coupled with the adaptive and numerical advantages of the local population, which together inhibit effective gene flow among populations (monopolization hypothesis; De Meester et al. 2002). Only a small proportion of the resting eggs hatch per wet period, resulting in a growing “resting egg bank” from which the subsequent generations are recruited (Brendonck and De Meester 2003). The resting egg bank acts as a buffer against invading genetic lineages. High dispersal rates, then, do not necessarily translate into high gene flow rates in these organisms (De Meester et al. 2002).
The evolution of species inhabiting the arid zone of Australia is closely coupled with the ongoing aridification, which started about 15 million years ago (mya) in the mid-Miocene (reviewed in Byrne et al. 2008 and Martin 2006). From then on to the end of the Pliocene (2.5 mya), arid zone taxa diversified, with most arid zone species dating to the Pliocene or even the mid-Miocene (Byrne et al. 2008). The transition from the Pliocene to the Pleistocene (2.5 mya) marked the initiation of glacial cycling, which corresponded in Australia with cold/dry glacial and warm/wet interglacial periods and an overall trend toward aridity. Within many arid zone species, the diversification of major genetic lineages occurred in the mid-Pleistocene (about 0.8 mya; reviewed in Byrne 2008), at which time shorter glacial cycles (∼40,000 years each) with lower amplitudes were replaced by longer cycles (∼100,000 years each) with higher amplitudes, coupled with an overall increase in aridity and an extensive expansion of arid environments (Byrne 2008).
The present study focuses on some of the most remarkable representatives of the Australian large branchiopods, those of the spinicaudatan taxon Limnadopsis. This taxon includes the world's largest clam shrimp L. birchii, which can reach up to 3 cm in length (Timms 2009). Limnadopsis species are restricted to temporary water bodies and never occur in permanent lakes or rivers. Limnadopsis is endemic in Australia, and its monophyly and Australian origin have been supported by recent molecular studies (Schwentner et al. 2009; Weeks et al. 2009). Limnadopsis birchii (Baird 1860) and L. tateiSpencer and Hall 1896 are found Australia-wide in arid and semiarid areas (Timms 2009). In an integrative taxonomic approach combining molecular and morpho-logical data, Schwentner et al. (2011) congruently delineated two species within L. tatei. One was identified as the true L. tateiSpencer and Hall 1896, which occurs in eastern, central, and western Australia (Schwentner et al. 2012), while the other, described as L. paratateiSchwentner et al. (2012), has so far only been reported from the catchment of the Paroo River in eastern Australia. Limnadopsis parvispinusHenry 1924 occurs from southern New South Wales to central and northern Queensland (Timms 2009; Schwentner et al. 2011). Schwentner et al. (2011) identified a clear genetic separation between the populations of the Murray–Darling Basin and the northeastern part of the Lake Eyre Basin, but suggested that they still represent a single species. (The lineages are referred to as L. parvispinus“Paroo” and “Buchanan,” respectively.)
In this paper, the phylogeography of four closely related Limnadopsis species is studied on the basis of two molecular markers (mitochondrial COI [cytochrome c oxidase subunit I] and nuclear ITS2 [internal transcribed spacer 2]) in order to assess the evolutionary and demographic history of species inhabiting Australian temporary water bodies. The main goal of this study is to identify the extent of population cohesion on various geographic scales (e.g., populations separated by a few kilometers versus those separated by more than 1000 km). Due to the patterns of genetic differentiation observed in populations of taxa with similar life-history traits in other parts of the world (e.g., other Branchiopoda with similar habitats, which are also dispersed via resting eggs), and those observed in species inhabiting isolated Australian freshwater systems (waterholes or groundwater springs), genetically differentiated populations are expected at geographic distances of much less than 100 km. Furthermore, a potential correlation existing between genetic structure and main drainage systems, as observed in many riverine species, is tested, because rare, severe floods may interconnect otherwise isolated temporary water bodies within river drainage systems, facilitating dispersal and gene flow within but not among drainage systems.