A Recent species of Frambocythere Colin, 1980 (Ostracoda, Crustacea) from a cave in South Korea; the first extant representative of a genus thought extinct since the Eocene

The non-marine ostracod genus Frambocythere Colin, 1980 (Limnocytheridae, Timiriaseviinae) had a previously known stratigraphical range from the Albian (Lower Cretaceous) through to the Lutetian (middle Eocene). However, during surveys of Seongryu Cave in Uljin-gun Province, South Korea, specimens of an extant species of Frambocythere were recovered. This extends the stratigraphic range of the genus by more than 40 million years to the present, and the species is, therefore, considered to be a relict ‘living fossil’. This newly described species is most similar in morphology to Frambocythere gr. tumiensis (Helmdach, 1978), reported from Thanetian (Paleocene) deposits of the Paris Basin. The living species was found in the hypogean realm, in contrast to the fossil species, which were all epigean. It is hypothesized that, like the genus Kovalevskiella Klein, 1963, which belongs to the same lineage, Frambocythere migrated from epigean to hypogean habitats.

diagnosis. Female: ventral and dorsal margins sub-parallel in lateral view, well-developed central sulcus, weakly developed anterior sulcus. Anterior margin rounded, posterior margin more truncated. Left valve with small caudal process and three stout, stubby spines on postero-ventral margin. 'Raspberry-type' micropustule ornamentation well developed. Male: slightly smaller than female, posterior less inflated than female. Hemipenis with large, straight, distal lobe with rounded tip, upper ramus hookshaped and small, lower ramus flattened with straight distal margin, rounded proximal margin. Copulatory process slender and simple.
description. Carapace (Fig. 2) length and height -see Table 1. Female -lateral view sub-rectangular, dorsal and ventral margins more or less parallel. Ventral margin slightly sinuous. Dorsal margin straight in anterior half, slightly convex in posterior half. Anterior margin equally rounded. Postero-dorsal margin angular, postero-ventral margin rounded. Two sub-parallel sulci running from dorsal margin downwards and slightly towards anterior margin, one positioned at mid-length, one shorter sulcus in anterior quarter (indicated with arrows on Fig. 2B). Surface of valves strongly ornamented with 'raspberry-type' micropustules. Left valve with small caudal process and three short, stubby spines on postero-ventral margin. Right valve overlaps left. Hinge lophodont, with cardinal teeth on left valve; anterior tooth gently curved and about 1.4 times length of posterior tooth. Dorsal view rounded posteriorly, triangular anteriorly, with strong indentation at midsulcus. Maximum width at posterior third. Male smaller than female with much less inflated posterior region. Left valve with one or two short, stubby spines. Dorsal view with maximum width just posterior of central sulcus.
Antennule (Fig. 3A) six segmented. First segment elongate, without setae. Second segment elongate with one long seta on ventral edge, and setules along dorsal edge. Third segment small, slightly elongate with one short, stout apical-dorsal seta. Fourth segment  longer, but thinner than third segment, with two apical-dorsal setae of differing lengths. Fifth segment longer than fourth, with three long and one shorter apical setae. Sixth segment small and elongate, with three long setae of differing lengths and long aesthetasc ya.
Antenna ( Fig. 3B) with four segments. First segment elongate, tapering slightly distally. Spinneret seta long and thin, reaching to end of apical claws. Second segment quadrate, with one, stout, long apical-ventral seta. Third segment very elongate, with  aesthetasc and seta on ventral edge, two setae on dorsal edge, and one short, stout apical-ventral seta. Fourth segment quadrate with three curved apical claws. Mandibular palp ( Fig. 3C) with four indistinct segments. First segment elongate, with branchial plate on outer edge and one stout, long seta on inner edge. Second segment quadrate, with one stout, long seta on inner edge. Third segment indistinctly divided from second segment, with two apical setae on outer edge and one apical seta on inner edge. Fourth segment short and small, indistinctly separated from third segment, with four apical setae. Mandibular coxa ( Fig. 3D) with seven teeth, and large, distally rounded setulous seta between two largest teeth. Small subapical seta on outer edge of coxa. Maxillula ( Fig. 3E) with elongate palp and three slender and longer endites, which gently curve towards inner edge. Palp with two curved, long, stout apical setae. Second and third endites both with four apical setae. First endite with two apical setae.
Fifth limb ( Fig. 3F) with four segments. First segment elongate with one long sub-proximal posterior seta, two setae on anterior edge and two apical-anterior setae. Second segment slender and elongate, with one short apical-anterior seta. Third and fourth segments approximately equal in length, fourth segment with short curved claw, with wide basal section, gradually tapering distally.
Sixth limb ( Fig. 3G) with four segments, longer than fifth limb. First segment elongate with one long sub-proximal posterior seta, two setae on anterior edge and one apical-anterior setae. Second segment slender and elongate, with one short apical-anterior seta.
Third and fourth segments approximately equal in length, fourth segment with short curved claw, with wide basal section, tapering gradually distally. Claw longer than claw of fifth limb.
Seventh limb (Fig. 3H) with four segments, all of which are more robust and wider than those of fifth and sixth limbs. First segment elongate with one long sub-proximal posterior seta, two setae on anterior edge and one apical-anterior seta. Second segment elongate, with one short apical-anterior seta. Third and fourth segments approximately equal in length. Claw of fourth segment long, mostly straight along length, with exception of distal tip; claw with subtriangular base, narrow middle section and thinner final section.
Rear of female body ( Fig. 4A) with finger-like caudal seta and numerous long, stout setules protruding below. Caudal ramus with two posterior, and one anterior stout, hirsute setae. Hemipenis ( Fig. 4B) with large basal capsule, with rounded outer edge, and straighter inner edge. Distal lobe large and elongate, not significantly curved, with evenly rounded tip. Upper ramus small, consisting of a rounded base tapering and curving distally to form a hook-shape. Lower ramus elongate with straight upper edge and curved lower edge, and with well-defined tip. Copulatory process slender, tapering gradually along length with blunt tip.
remarks. The carapace shape of Frambocythere relicta n. sp. is most similar to that of Frambocythere gr. tumiensis (Helmdach, 1978) reported from the Thanetian (upper Paleocene) of the Paris Basin (Ducasse et al., 1985). Both species have a weakly developed anterior sulcus, and the general lateral outline and ornamentation are very similar. Frambocythere relicta n. sp. is slightly less elongate than Frambocythere gr. tumiensis, and the latter has only very small spines on the postero-ventral margin of the left valve.
Frambocythere valeroni Tambareau, 1991(in Tambareau et al., 1991 from the Ypresian (early Eocene) of SW France is also similar to Frambocythere relicta n. sp.; both F. valeroni and Frambocythere relicta n. sp. have a weakly developed anterior sulcus. However, F. valeroni is more elongate than Frambocythere relicta n. sp. with a maximum height in lateral view anterior of mid-length.

dIscussIon hypogean ostracod Fauna of south Korea
Previous surveys of seven limestone caves during the 1960s in the northeastern part of South Korea revealed three hypogean species/ subspecies of ostracods: Pseudocandona morimotoi (McKenzie, 1972), Cavernocypris coreana coreana (McKenzie, 1972) and Cavernocypris coreana elongata (McKenzie, 1972). All three taxa are potentially endemic to the Korean Peninsula, although at least one, C. coreana elongata, is not restricted to cave environments as it was later found in spring runoffs at the surface (Chang et al., 2012). Frambocythere relicta n. sp. increases the known cave ostracod fauna of South Korea to four species, and is the first representative of the superfamily Cytheroidea in the hypogean realm of South Korea.
The genus Kovalevskiella is similar to Frambocythere, but differs from it in that it has only one sulcus on each valve. It is known from lacustrine deposits in the late Oligocene and Miocene of Western Europe and the Pliocene of the Paratethys (Carbonel et al., 1986). Living representatives are now restricted to the hypogean realm of central and southeastern Europe, and Turkey. Males have not been reported for any of the species in the genus and it is considered to be entirely parthenogenetic. The genera Kovalevskiella and Frambocythere are considered to be more closely related to each other than to other genera in the group (Gidó et al., 2007;Colin, 2011).
The anterior sulcus of Frambocythere, the feature that separates the genus from Kovalevskiella, is rather weakly developed in F. relicta n. sp. and two of the younger fossil forms, F. gr. tumi ensis, and F. valeroni, compared with other fossil species. The anterior sulcus is probably a plesiomorphic character within the Frambocythere-Kovalevskiella lineage, which became reduced and eventually lost in some taxa, giving rise to the genus Kovalevskiella. The difference between Frambocythere species with a reduced anterior sulcus and Kovalevskiella species is, therefore, rather small. However, the anterior sulcus is a persistent character in numerous species/subspecies of Frambocythere, and has a good fossil record dating back to the Cretaceous, long before species without it, i.e. Kovalevskiella spp., appeared. Its presence, even when weakly developed (and its absence in Kovalevskiella), can therefore be used to help identify phylogenetic lineages within the group, and so we consider it to be a useful generic character.
comparison of Frambocythere relicta n. sp. with living Kovalevskiella ssp The living Korean Frambocythere species provides an opportunity to compare the appendages of these two genera for the first time.
Of the five extant Kovalevskiella species, the antennule of Kovalevskiella cvetkovi (Danielopol, 1969) is most similar to that of Frambocythere relicta n. sp.; other species either have fewer antennule segments (Kovalevskiella rudjakovi (Danielopol, 1969)) or one fewer apical seta on the fourth antennule segment (Kovalevskiella phreaticola (Danielopol, 1965), Kovalevskiella bulgarica (Danielopol, 1970) and Kovalevskiella dani Karanovic, 2003). The antenna of Frambocythere relicta n. sp. most closely resembles that of K. bulgarica, as other Kovalevskiella species have one fewer seta on the anterior margin of the second endopodal segment, including K. cvetkovi. The mandibular palp of Frambocythere relicta n. sp. is similar to those of three Kovalevskiella species, K. phreaticola, K. cvetkovi and K. bulga rica; the mandibular palps of K. dani and K. rudjakovi have fewer setae on the second and third segments. The other appendages of Frambocythere and Kovalevskiella are very similar, although K. phreaticola appears to have only one seta on the dorsal margin of the first segments of the sixth and seventh limbs (Danielopol, 1965) (two setae in Frambocythere relicta n. sp. and other Kovalevskiella species). As no males of Kovalevskiella are known, a comparison of the male sexual organ of Frambocythere cannot be made. However, the general structure of the male sexual organ does resemble those of the other two extant Timiriaseviinae genera, Metacypris Brady & Robertson, 1870 and Dolekiella Gidó et al., 2007. Overall, all features of the female appendages of Fram bocythere relicta n. sp. appear in at least one extant Kovalevskiella species, confirming that these two genera are very closely related. Additionally, all features in Kovalevskiella appear in Fram bocythere relicta n. sp.; thus, Frambocythere is not excluded from being the ancestor of Kovalevskiella. relict species or atavistic characters? There are two scenarios that could potentially explain the presence of an extant Frambocythere species after the apparent extinction of all other species of the genus. The first is that it is a true survivor of the genus Frambocythere and, while other species of the genus succumbed to extinction tens of millions of years ago, a lineage survived to the present day. The other scenario is that its anterior sulcus of the carapace, the morphological feature that separates Frambocythere from the genus Kovalevskiella, is an atavistic feature that has resurfaced in the extant Kovalevskiella. This would require that Frambocythere is the ancestor of Kovalevskiella, which a detailed analysis of the appendages does not exclude (see above). It would also require the species to have reverted from asexual to sexual reproduction, as all Kovalevskiella species, both fossil and extant, are parthenogenetic. Such a reversion to sexual reproduction would also be an atavistic feature, and is something that has not been recorded in ostracods. Two atavistic features occurring in the same species (the re-emergence of the anterior sulcus and the reversion to sexual reproduction) is considered to be unlikely, and so we favour the first scenario, i.e. Frambocythere relicta n. sp. is a relict species of the genus, and not a Kovalevskiella species with atavistic features. Molecular comparisons of Frambocythere relicta n. sp. with extant Kovalevskiella species could test this hypothesis. Danielopol (1970; hypothesized that the hypogean species of the Kovalevskiella-lineage living today in central and southeastern Europe, and Turkey are relicts of an ostracod fauna that lived in surface waters during the Tertiary. While some surface species became extinct, others migrated to the hypogean realm; Kovalevskiella is postulated to have penetrated the subterranean realm between the Lower Oligocene and Lower Pleistocene (Danielopol, 1980). Carbonel et al. (1986) further suggested that this migration to the hypogean realm was facilitated by their morphology, broad ecological ranges and parthenogenetic mode of reproduction. We hypothesize that a similar scenario may have occurred in the genus Frambocythere; it too was originally widely distributed and long ranging, and was an epigean taxon, but today, it is apparently restricted to Korea in a hypogean habitat. However, Frambocythere relicta n. sp. is a sexual species, suggesting that at least in this case, parthenogenesis was not an important pre-adaptation for colonizing the hypogean realm.

From the epigean to hypogean realm
The long gap (approximately 40 Ma) between the youngest fossil Frambocythere known (an epigean species) and the living hypogean specimens could indicate that Frambocythere entered the hypogean realm an extremely long time ago, and thus 'disappeared' from the epigean fossil record. However, data on fossil freshwater ostracod taxa in eastern Asia are scarce, and so we cannot rule out that younger epigean Frambocythere species existed in this region. We are, therefore, unable at present to estimate the timing of the colonization of the hypogean realm by the genus to any satisfactory degree.

acKnowLedGeMents
We thank the Natural Heritage Division of the Cultural Heritage Administration and the Office of Seongryu Cave, Ulijin County, Republic of Korea for their help with this study. We also thank Dan Danielopol (University of Graz) and Dave Horne (University of London) for their useful reviews and comments. This work was partly supported by the project of Discovery of Indigenous Species from Korea sponsored by the National Institute of Biological Resources (NIBR) under the Ministry of Environment, Korea.

Manuscript received 1 december 2011 Manuscript accepted 13 January 2012
Scientific Editing by Alan Lord reFerences