Introduction

Twelve freshwater podocopid ostracods belonging to the superfamily Cypridoidea and family Cyprididae are known from Tasmania (Table 1). The family Candonidae and representatives of the superfamily Cytheroidea, chiefly the family Limnocytheridae, have been recorded from Quaternary deposits (De Deckker 1982b) from Pulbeena and Mowbray swamps in north-west Tasmania. Darwinulidae are represented in Tasmania by one unnamed species of Penthesilenula Rossetti and Martens, 1998 (see Martens and Rossetti 2002) from the same Quaternary site.

Since their first appearance in the early Palaeozoic, podocopid ostracods have occurred in both marine and freshwater habitats, and today can be found from abyssal depths to small freshwater puddles, and even terrestrial environments (Pinto et al. 2003, 2004, 2005, 2008; Martens et al. 2004). All three major podocopid lineages, Cypridocopina, Darwinulocopina and Cytherocopina, have several representatives living in terrestrial habitats. Although Darwinulidae has no species restricted exclusively to terrestrial habitats, several species of the genera Vestalenula Rossetti and Martens, 1998 and Penthesilenula Rossetti and Martens, 1998 have been collected from terrestrial environments in Brazil (Pinto et al. 2003, 2004, 2005) (for the review of “semi-terrestrial ostracods” see Appendix).

Schornikov (1980) argued that ostracods found in terrestrial environments cannot be considered true terrestrial animals, as they were always covered by water, even if only a drop. This claim was rejected by Pinto et al. (2005) and Martens et al. (2004), who considered “terrestrial” all animals living outside free standing or flowing water. Danielopol and Vespremenau (1964) observed Scottia pseudobrowniana (Jones, 1850) move towards more humid areas, but closed its valves and stopped moving in non-saturated conditions, only becoming active again when in contact with water. Individuals of Scottia audax (Chapman, 1961)tended to move towards more damp parts of the litter as the available water decreased (Chapman 1961). Horne et al. (2004) recorded strong hydrotaxis in Terrestricythere elisabethae Horne, Smith, Whittaker and Murray, 2004, and death if desiccated for more than 10 minutes. The ability to inhabit environments such as leaf litter and other damp habitats may be advantageous, since it can allow a species to widen its area of distribution significantly and avoid predators occurring in fully aquatic habitats. It is easy to imagine how these tiny animals move, and one direct example is the finding of Microdarwinula zimmeri (Menzel, 1916) and Scottia pseudobrowniana on the floating fen soil (always in the part of the fen covered with water) in Romania (Danielopol and Vespremenau 1964). On the other hand, this way of life has its obvious limits, since such places are often subject to rapid desiccation. One more argument for not considering these ostracods truly adapted to the terrestrial environment is the fact that they are actually not widely distributed as would be expected with this adaptation. We know of only three exceptions where significantly broader distribution ranges, including trans-continental, are implied: Mesocypris pubescens Daday, 1910 has been recorded from Mount Kilimanjaro (Daday 1910), Kenya (Klie 1939) and South Africa (Lawrence 1953); Scottia audax from New Zealand (Chapman 1961) and widely throughout eastern Australia (De Deckker 1980, 1983); and Scottia pseudobrowniana from Europe, Asia (see Meisch 2000) and North America (Cole 1966, Külköylüoglu and Vinyard 2000). The broad distribution of these three species is discussed further in this paper and it is shown that they actually comprise additional unrecognized species. Redescriptions of these species is beyond the scope of this paper, but “species” as recognized herein are included in the key to species of Scottiinae presented at the end of this paper.

Animals living away from free-standing or flowing water have usually very hirsute shell and soft parts, which probably allows them to retain enough moisture in semi-terrestrial surroundings (Horne et al. 2004). Pinto et al. (2005) also consider parthenogenetic reproduction, commonly found in these ostracods, as one of the pre-adaptations to this environment. Although Chapman (1961) did not provide a detailed taxonomic description of Mytilocypris audax, she gave a very interesting observation on the behaviour of these animals. She suspected that the juveniles hatch inside mother's shell, since she observed large eggs within the shell with the outline of valves showing. Even more, in a live culture, in which young specimens were suddenly noticed, no laid eggs were seen. All this may suggest brooding care, never noticed before in Cypridoidea.

During the recent study of Tasmanian caves, two interesting new species were collected. One is named and described here, while the other is only briefly described but left as an open nomenclature, because only a juvenile specimen was collected. They both belong to the genus Austromesocypris Martens, De Deckker and Rossetti, 2004, a very peculiar and otherwise entirely “terrestrial” genus known only from Australia (Martens et al. 2004).

Study area

Precipitous Bluff is located near the south coast of Tasmania in a remote and inaccessible part of the Tasmanian Wilderness World Heritage Area (Figure 1). The area remains essentially undisturbed by human activities except for occasional visits by bushwalkers and cavers. An extensive deposit of highly karstic and cavernous limestone of Ordovician age outcrops over about 10 km² between 0–300 meters above sea level (asl) on the western and southwestern flanks of Precipitous Bluff (1120 m asl) (Hughes 1957; Dixon and Sharples 1986) (Figure 2). The study area lies within the per-humid precipitation effectiveness province of Gentilli (1972). The nearest meteorological station is at Hastings located 25 km east-northeast, where mean annual rainfall is 1, 416 mm. The cool, moist climate supports a dense vegetation cover of wet sclerophyll and cool temperate rainforest (Reid et al. 1999) (Figure 3).

The caves have developed by allogenic recharge from streams descending from the upper slopes of the mountain which sink underground on contact with the limestone. Autogenic recharge is also received from rainfall directly on the limestone outcrop, which percolates downward via a multitude of vertical shafts and fissures in the limestone (Figure 4). The descending subsurface flow paths coalesce at lower levels to form an integrated dendritic cave drainage network which discharges from resurgence caves situated at base level on the slope-plain juncture. Damper Cave is a major resurgence cave with more than one km of surveyed underground passages (Figure 5). The cave consists of a major conduit carrying a permanent flowing stream with a bedrock-gravel-sandy bed, in addition to numerous small tributary streams and seepages.

The cave fauna at Precipitous Bluff was surveyed as part of a state-wide biodiversity survey of Tasmanian karsts undertaken by Eberhard et al. (1991). This study discovered that the Precipitous Bluff karst supported the richest assemblage of locally endemic and obligate cave-dwelling species in Tasmania (Eberhard 1996). Initially, most of this assemblage comprised terrestrial cave obligate species (troglobites) including described species of Coleoptera (Eberhard and Giachino 2011; Moore 1978), Diplopoda (Mesibov 2005), Opiliones (Hunt 1990; Hunt and Hickman 1993), synotaxid (Forster et al. 1990) and micropholcommatid (Rix and Harvey 2010) spiders, in addition to other undescribed species of Opiliones, Araneae, Pseudoscorpionida, Oniscidea, and Collembola (Eberhard et al. 1991). A significant fauna of obligate aquatic subterranean species (stygobites) in the Precipitous Bluff caves was also indicated by the discovery of Australia's first stygobitic gastropod Pseudotricula eberhardi Ponder 1992, followed by the description of an additional seven species of cave dwelling hydrobiid gastropods (Ponder et al. 2005). Besides the hydrobiids, the sampled aquatic macrofauna included mayflies (Ephemeroptera) and crustaceans (Amphipoda and Syncarida), the latter possessing clear stygomorphies (loss of eyes, pigment and elongation of appendages) indicating their status as stygobites. The initial survey by Eberhard et al. (1991) focused on collecting macro-invertebrates, but Ponder et al. (2005) concluded that additional sampling, especially targeted towards aquatic micro- and meiofauna, was likely to further increase the richness of subterranean species known from the Precipitous Bluff karst. This possibility provided the impetus for the current study.

Field methods

Aquatic micro- and meiofauna were sampled from a variety of habitats in Damper Cave in May 2011 using 150 µm mesh plankton nets. Three types of aquatic habitats were sampled: (1) main stream (Site A); (2) small tributary stream (Site B); and (3) seepage water from dripping stalactites (Sites, C, D, E). To sample the main stream benthos, shallow interstitial and stream drift fauna, two 300 mm diameter plankton nets were installed in the main flow of the stream approximately 30 m upstream from the resurgence entrance (Figure 6). Benthic and shallow interstitial fauna was sampled by ‘kick' sampling which involved walking upstream from the nets for a distance of approximately 50 m while ‘kicking' the gravel-sandy sediments to dislodge fauna which was swept downstream into the nets. A small tributary stream approximately 20 m inside the resurgence entrance was sampled by placing a bucket and net underneath a vertical trickle (Figure 7). Seepage water from dripping stalactites was sampled in a similar manner. The nets were left in the cave for two days and then recovered, where after the net contents were elutriated and preserved in 100% ethanol.

Taxonomic methods

Specimens were dissected and mounted on microscope glass slides in Faure's medium, which was prepared following the procedure of Stock and von Vaupel Klein (1996). The dissected appendages were then covered with a coverslip and observed under the Leica L2 stereomicroscope and Leica DM 2500 compound microscope with N-plan objectives, and the drawings were made using a drawing tube. The photographs of ostracod shells were taken with a Leica M205 C microscope fitted with a Leica DFC420 digital camera, and montaged using the Leica Application Suite V3 software. All material is deposited at the Tasmanian Museum and Art Gallery (TMAG), Hobart.

Terminology. In the present paper the terminology for the most posterior appendage on the body, the so-called uropodal ramus, follows Meisch (2007). Terminology of other soft parts follows Broodbakker and Danielopol (1982), Martens (1987), and Meisch (1996).

Abbreviations used in text and figures. A1-antennula; A2-antenna; LV-left valve; L5-L7-fifth-seventh limb; Md-mandibula; Mxl-maxillula; RV-right valve; UR-uropodal ramus.

Taxonomy Class Ostracoda Latreille, 1806
Subclass Podocopa Müller, 1894
Order Podocopida Sars, 1866
Superfamily Cypridoidea Baird, 1845
Family Cyprididae Baird, 1845
Subfamily Scottiinae Bronstein, 1947 Discussion Ecology

No Ostracoda have previously been described from Tasmanian caves or other groundwater environments. Illife (1988) reported a collection of ostracods from Honeycomb Cave at Mole Creek in northern Tasmania; however these remain unidentified. As already noted, biospeleological collections to date have targeted macro-invertebrates while aquatic micro- and meio-fauna have been poorly investigated (Ponder et al. 2005). The absence of Candoninae in Tasmanian caves may be an artefact of poor sampling effort, but it is also possible that this group is substituted here by Austromesocypris Martens, De Deckker & Rossetti, 2004 species that have occupied a new habitat through the lack of competition with Candoninae, the dominant subterranean ostracod group elsewhere in the world (Karanovic 2007). Despite a substantial sampling effort during this study, only one complete specimen of each species has been collected. Subterranean ostracods are often rare in subterranean samples (Karanovic 2007, Karanovic and McKay 2010).

Discovery of cave dwelling ostracods described in this paper is of interest for two reasons. Firstly, members of Cyprididae are rarely found in subterranean waters. Secondly, the genus Austromesocypris is otherwise known to comprise entirely “terrestrial and semi-terrestrial” species. The rare finds of Cyprididae in subterranean waters are mostly records of morphologically unmodified taxa which occur as facultative or incidental inhabitants of groundwaters. The only documented case of an obligate association of cyprids with groundwaters involves the genus Pseudocypridopsis Karanovic, 1999, with two described stygobitic species from the Balkan Peninsula (Karanovic 1999, 2000). In Western Australia, where a great deal of sampling in wells and water bores has revealed a globally significant richness of stygobitic Crustacea (Eberhard et al. 2005, 2009), no truly stygobitic Cyprididae occur, although unmodified epigean species are not infrequently collected especially in wells which are open to the surface and therefore easily colonized by epigean species.

The collection of a whole specimen of Austromesocypris bluffensis sp. n. in the main stream implies that this animal could have been living in the benthos or interstitial of the main stream, but also does not preclude an origin from a tributary stream or seepage waters, of which the latter two ultimately discharge into the main stream. The collection of valves in seepage water from dripping stalactites and a complete specimen of Austromesocypris sp. in the small tributary stream confirms the occurrence of Austromesocypris in these habitats also. Notwithstanding they might still have originated from the surface which lies only a few tens of meters above Damper Cave at the point where the collections were made. Certainly the cool temperate rainforest represents a near-permanently moist habitat for hygrophilous invertebrates including, potentially, terrestrial/semi-terrestrial ostracods. The highly permeable karst at Precipitous Bluff ensures a close eco-hydrological connection between epigean and hypogean environments, confirmed by the occurrence of other normally epigean-terrestrial invertebrates such as charopid gastropods accidentally washed into the caves by percolating seepage waters (S. Eberhard personal observation).

Adaptive morphology

Some adaptive morphologies, typically comprising numerous reductions, displayed in semi-terrestrial epigean ostracods are similar to those seen in aquatic subterranean ostracods as already pointed out by several authors (Danielopol 1978, Danielopol and Betsch 1980, Pinto et al. 2005, 2008). Danielopol and Betsch (1980) also point out the close phylogenetic relationships between the semi-terrestrial Candoninae from Madagascar and their interstitial aquatic relatives. This may additionally support the view of Karanovic (2007) that the subterranean Candoninae genus Nannocandona Eckman, 1914 should belong to Terrestricandonini instead of Candonini because of the many morphological characters shared with members of the former tribe. It is also worth noting that the subfamily Timiriaseviinaenot only contains the only freshwater semi-terrestrial cytheroid ostracod Intrepidocythere Pinto, Rocha and Martens, 2008 but also is the only cytheroid subfamily with true stygobiont species such as those in the genera Kovalevskiella Klein, 1963 and Gomphodella De Deckker, 1981.

The unusual shape of the two new Tasmanian species, one with trapezoidal shape, the other with triangular shape, suggests that these might also be true stygobites as such modification is very common among subterranean Candoninae (Danielopol 1978, Karanovic 2007). In addition to this potentially convergent stygomorphic character, typical for Ostracoda, the Austromesocypris from Precipitous Bluff also exhibit classic convergent regressive stygomorphic characters seen in subterranean Crustacea, including depigmentation, loss of eyes and reduction in body size. On the other hand, these Tasmanian specimens do not exhibit any other obvious stygomorphic characters such as elongation of appendages as seen in Candoninae and most other stygomorphic Crustacea. It should be noted that the dominantly terrestrial/semi-terrestrial group Scottiinae typically exhibit reduction in body size and size of segments, a gracile appearance, and anopthalmia.

Systematics

Austromesocypris is according to Martens et al. (2004) most closely related to Mesocypris Daday, 1910 and they share a reduced number of free segments on the A1, fused second and third endopodal segments on the L6, a short h3 seta on the terminal segment of the same appendage, and asymmetrical caudal rami. Mesocypris is a paraphyletic taxon according to the cladistic tree obtained by Martens et al. (2004), and indeed many species deviate from the characters that should define the tribe Mesocypridini to which the two genera belong according to Martens et al. (2004). For example, at least two species have the second and third endopodal segments partly separated on the L6: Mytilocypris terrestris Harding, 1953 and Mytilocypris pauliani Danielopol & Betsch, 1980. The h3 seta is longer in all Mesocypris species than in Austromesocypris and its length is somewhere in between for the latter genus and Scottia Brady & Norman, 1889. Martens et al. (2004) stated in their revised diagnosis that in tribe Mesocypridini h3 is short, while in Scottiini it is claw-like. One of the Madagascar species, Mytilocypris pauliani Danielopol & Betsch, 1980 actually has a claw-like h3 seta. Another doubtful character of the tribe Mesocypridini is the morphology of the UR. While the posterior margins of the right and left ramus are indeed asymmetrical (one hirsute, other with teeth or spines) in all previously known species, only Austromesocypris australiensis has one posterior seta transformed into a strong claw and the other seta-like. In all other Mesocypridini species this difference is much less obvious, especially in other species of the genus Austromesocypris.

All species of Austromesocypris have only two swimming setae on the A2, while in Mesocypris the number varies from six to two. There are, however, two characters which indicate the two genera are more closely related to each other than either is to the genus Scottia. Namely, the terminal segment on the Md-palp is much more elongate and the number of setae on the first segment of the Mxl-palp is reduced in Mesocypris and Austromesocypris in comparison to Scottia. According to Chapman's (1961) illustration of the Md of Scottia audax (Chapman, 1961)and Chapman's (1963) drawings of the Mxl of Scottia insularis Chapman, 1963, the two New Zealand species appear to be more closely related to members of Mesocypridini. Scottia audax was originally described in the genus Mesocypris, but De Deckker (1980) transferred it to the genus Scottia because of the morphology of the L6, namely the presence of the long, claw-like h3 seta, and at the same time also suggested that Scottia insularis Chapman, 1963 might belong to Mesocypris due to the short h3 seta. Both New Zealand species were retained in the genus Scottia by Martens et al. (2004), but those authors pointed out the unusual distribution of the genus, with two species, Scottia pseudobrowniana Kempf, 1971 widely distributed in the Holarctic (Meisch 2000) and Scottia birigida Smith, Matzke-Karasz, Kamiya & Ikeda, 2002 endemic to Japan (Smith et al. 2002).

The position of the two New Zealand species is not certain at the moment, but they definitely should be excluded from the genus Scottia. While they share many characters with Mesocypris, most of which are variable, they also have a reduced number of segments on the A1 like Austromesocypris. Fusion of A1 segments can also be partial, as noted by Matzke-Karasz (1995) for Scottia audax. Namely, borders between segments 3+4 and 5+6 are incompletely formed and only visible on the interior side, a fact noticed by Martens et al. (2004) as well. Another character that may support a closer relationship of Scottia audax to Mesocypridini is the reduced number of rays on the exopod of L5. Namely, both Palearctic Scottia species have six rays, while the New Zealand species has only four (see drawings of Matzke-Karasz 1995). The number is further reduced to two or three setae in Mesocypris and Austromesocypris. This character needs to be taken with caution, because the exopod of L5 may be easily damaged during slide preparation and the setae easily fall off.

Misidentified species

In our opinion Australian records of Scottia audax (Chapman, 1961) by De Deckker (1980) are incorrect, that material representing a different, as yet undescribed species. This is even more apparent from the drawings of Scottia audax provided by Matzke-Karasz (1995) who studied De Deckker's (1980) material. The most significant difference between the Australian species and Scottia audax is the morphology of the UR. The posterior seta is longer, stronger and more closely situated to the distal end of the ramus in the Australian species, but the claws are stronger and the posterior margin of the ramus is covered with small teeth (or at least thick spines) in Scottia audax. Another finer detail is the appearance of the exopodal setae of the A2, which in the Australian species has one long and two short, subequal setae whereas in Scottia audax the two shorter setae are unequal in length, with one being at least twice as long as the other (as in Austromesocypris bluffensis). The chaetotaxy of the Md-palp is also somewhat different, as all the dorsal setae in Scottia audax are longer than those of the Australian species. It is interesting that Scottia audax and the Australian species have the S2 seta turned the same way as in Austromesocypris bluffensis sp. n. and that all three species also have four setae postero-distally on the penultimate segment.

In our opinion De Deckker's (1983) report of Austromesocypris australiensis also represents two species which differ most significantly in the appearance of the hemipenis, prehensile palps and the number of the rosettes on the Zenker organ. One was found in far north Queensland, and the other in New South Wales and Victoria, but both were identified as Austromesocypris australiensis by De Deckker (1983). Previous collections of Austromesocypris tasmaniensis may also contain two species, as De Deckker (1983) observed some differences in the length of the shell and morphology of the UR between specimens taken from two localities. These samples need to be studied in more detail to draw an accurate conclusion.

Misidentified Scottiinae are not confined to Australian records. Reports of the “widely distributed” Scottia pseudobrowniana from Tennessee (Cole 1966) and Nevada (Külköylüoglu and Vinyard 2000) are also probably incorrect. The specimens from Tennessee (east) resemble Scottia pseudobrowniana more than those from Nevada (west). While the Nevada material has an almost straight posterior margin of the shell, the one from Tennessee has a rounded posterior margin, similar to that of Scottia pseudobrowniana, but with both ends being narrower than in the latter species. Specimens of the North American Scottia pseudobrowniana have shorter swimming setae on the A2 than the European Scottia pseudobrowniana. Indeed the setae in the Nevada material are so short that they only slightly overpass the first endopodal segment. Specimens of Scottia pseudobrowniana from the three separate geographic locations differ mostly in the morphology of the hemipenis. Although neither Cole (1966) nor Külköylüoglu and Vinyard (2000) labelled or drew this organ in detail, the outline they provide indicates enough differences to claim that three separate species are represented. Scottia pseudobrowniana has all segments on the A1 free, while Külköylüoglu and Vinyard (2000) illustrated a 6-segmented A1. If this is correct, it would represent an outstanding character. It must be noted, however, that the illustrations of Külköylüoglu and Vinyard (2000) show limited detail and may not correctly show the relevant characters.

Scottia pseudobrowniana from Russia, as illustrated by Bronstein (1947), is very similar to the figures of the Nevada species of Külköylüoglu and Vinyard (2000), but they differ in the length of the swimming setae (longer in the Russian species). Bronstein's (1947) illustration of the hemipenis is also rather similar to the figures given by Külköylüoglu and Vinyard (2000). Smith et al. (2002) also noted that Scottia pseudobrowniana reported from Russia and Nevada are quite different from the European populations, and suggested that the respective material needs to be re-identified. We conclude that the Russian and both North American records of Scottia pseudobrowniana represent new species and that the differences cannot be attributed to variability. This is supported by several reports of the absence of morphological variability among European populations of Scottia pseudobrowniana (see Kempf 1971, Danielopol and Vespremenau 1964, Petkovski 1966, Matzke-Karasz 1995, Meisch 2000).

Key to genera and species of Scottiinae