Forever in the dark: the cave-dwelling azooxanthellate reef coral Leptoseris troglodyta sp. n. (Scleractinia, Agariciidae)

Abstract The coral species Leptoseris troglodyta sp. n. (Scleractinia, Agariciidae) is described as new to science. It is the first known azooxanthellate shallow-water agariciid and is recorded from the ceilings of caves at 5-35 m depth in West Pacific coral reefs. The corals have monocentric cup-shaped calices. They may become colonial through extramural budding from the basal coenosteum, which may cause adjacent calices to fuse. The size, shape and habitat of Leptoseris troglodyta are unique compared to other Leptoseris species, many of which have been recorded from mesophotic depths. The absence of zooxanthellae indicates that it may survive well in darkness, but endolithic algae in some corals indicate that they may be able to get some light. The presence of menianes on the septal sides, which may help to absorb light at greater depths in zooxanthellate corals, have no obvious adaptive relevance in the new species and could have been inherited from ancestral species that perhaps were zooxanthellate. The new species may be azooxanthellate as derived through the loss of zooxanthellae, which would be a reversal in Leptoseris phylogeny.


introduction
Reef-dwelling species of the genus Leptoseris Milne-Edwards and Haime, 1849 (Scleractinia: Agariciidae) consist of foliaceous corals that are common in poorly illuminated environments, such as the deepest parts of reef slopes and vertical rocky walls with crevices, caves, tunnels and overhangs (Dinesen 1980(Dinesen , 1982(Dinesen , 1983Waheed and Hoeksema in press). They are considered skiophilous (shade-loving) or cavernicolous, i.e., living in caves (Dinesen 1982(Dinesen , 1983. Because they appear to show more preference for dark habitats than many other reef corals, Leptoseris species constitute an important component of zooxanthellate scleractinian coral communities at mesophotic depths (30-150 m) Kelley 2007, Chan et al. 2009, Bare et al. 2010, Bongaerts et al. 2010, Kahng et al. 2010, Dinesen et al. 2012. They may even occur deeper, with records of 153 and 165 m by Leptoseris hawaiiensis Vaughan, 1907, in the Pacific Ocean Jokiel 1986, Kahng andMaragos 2006), and 145 m by L. fragilis Milne Edwards and Haime, 1849, in the Red Sea (Fricke and Schuhmacher 1983, Fricke et al. 1987, Kaiser et al. 1993. Because some Leptoseris species inhabit deep and poorly accessible habitats, they may not all be well known. An example is the recently discovered L. kalayaanensis Licuanan & Aliño, 2009, which so far has only been recorded from rocky substrates at 13-28 m depth in the South China Sea basin (Licuanan andAliño 2009, Hoeksema et al. 2010). It shows a distinct brown and white coloured pattern on its upper surface, consisting of areas that are either with or without zooxanthellae.
The Agariciidae were not known to include true deep-sea species but according to recent phylogenetic studies, the solitary attached deep-water coral Dactylotrochus cervicornis (Moseley, 1881), which was originally classified with the Caryophylliidae, is also a member of the Agariciidae (Kitahara et al. 2010. This species has a recorded depth range of 73-852 m, is therefore considered ahermatypic and probably azooxanthellate (for terminology see Zibrowius 1985, Cairns and. It is monocentric and has smooth-edged septa that bear 2-5 elongate ridges (menianes or latera), which are considered characteristic for the Agariciidae (Kitahara et al. 2010. Because D. cervicornis is predomintly from deep water, it is considered the first known extant azooxanthellate agariciid. D. cervicornis holds a basal position in a recent phylogeny reconstruction of extant Agariciidae and because extinct solitary agariciids from the Middle Cretaceous were also solitary, it is assumed that the ancestor of the Agariciidae, which nowadays predominantly consist of colonial and zooxanthellate species, was also solitary and azooxanthellate ).
In the present paper a new agariciid coral species is described that is entirely azooxanthellate and dwells on ceilings of caves in steep reef slopes and walls. No co-occurrence with any zooxanthellate scleractinians was observed. Although its calices are relatively small, cup-shaped and predominantly monocentric, it resembles species of Leptoseris, which otherwise is known to consist of zooxanthellate species with polycentric ("circumoral") calices (Wells 1956). It is furthermore unique among extant reef-dwelling Agariciidae because it is modular (colonial) through extramural budding by growing new calices from a basal coenosteum, which eventually may fuse. It has been found in the western Pacific, including eastern Indonesia, central Philippines, Papua New Guinea, Palau and Guam. Most of its presently known distribution range overlaps with the centre of maximum marine species richness, the so-called Coral Triangle (Hoeksema 2007).

Methods
Specimens were observed, photographed and collected while diving with the help of SCUBA. All specimens were encountered below 5 m depth on the ceilings of caves inside steep reef slopes and walls, usually in areas with limestone outcrops. The caves measured one to several meters in width and several meters in length, enabling easy access and maneuvering for observations. Use of an underwater torch was indispensable to locate the corals. Collected specimens were soaked in fresh water or in sodium hypochlorite solution for cleaning. They were deposited in the Coelenterata collection (RMNH Coel.) of Naturalis Biodiversity Center, Leiden (formerly known as Rijksmuseum van Natuurlijke Historie). Specimens from Cebu were already available in the RMNH collection before they were photographed and collected for the present research. SEM photographs were taken with a Jeol 6480LV electron microscope operated at 10 kV.          2000; Figure 2f ).
Etymology. The epithet troglodyta (noun) means cave dweller in Latin, derived from ancient Greek for "one who dwells in holes".
Distribution. Records are from coral reefs, usually in areas with limestone outcrops: Indonesia (East Kalimantan, North Sulawesi, Southeast Sulawesi), the central Philippines (Cebu, Bohol), Palau, eastern Papua New Guinea, and the Marianas (Guam) (Figure 9).

Systematics
Leptoseris troglodyta sp. n. has a habitat and growth form unlike any other known Leptoseris; its corals are cavernicolous and azooxanthellate, and have small, monocentric calices that may multiply by extramural budding and fuse. Other Leptoseris species are polycentric by circumoral and marginal budding (Dinesen 1980, Veron and Pichon 1980, Veron 2000, Licuanan and Aliño 2009. All congeneric species are larger, which may be related to their symbiosis with zooxanthellae. Polycentric species are usually bigger than close relatives with monocentric calices, as demonstrated for mushroom corals (Hoeksema 1991b, Gittenberger et al. 2011.
The new species lacks pigments of its own, like many cavernicolous (= troglobiotic) animals. Although there are no zooxanthellae in its soft tissue, it usually contains green or red shade-adapted endolithic cyanobacteria imbedded in the skeleton, which have also been reported from Leptoseris fragilis Milne-Edwards and Haime, 1849 (Schlichter et al. 1997). Dinesen (1983) mentions the occurrence of "numerous Leptoseris cf. scabra (G. Hodgson, M. Ross, pers. comm.)", which were observed in 1981 on the ceiling inside the large Marigondon Cave (Mactan Island, Cebu), the Philippines. They were found "further back in the cave, in gloomier conditions" than the cave entrance at ca 30 m depth. It is likely that this record pertains to specimens of the new species. Two corals of the present study were collected at that site in 1981 and were available for study in the RMNH collection in Leiden. These specimens confirm that the new species was present at that locality at that time. Museum collections may indeed help to retrieve information on the earlier occurrence of coral species that have not been recorded before ( Van der Meij et al. 2010, Van der Meij and Visser 2011. However, without field observations it would not have been possible to know that the new species lacks zooxanthellae. Despite its unique small monocentric corallites and lack of zooxanthellae, the new species is classified with Leptoseris because of similarly shaped septa and costae, and by its thin, cup-shaped or foliaceous coralla somewhat resembling those of L. fragilis. The latter species features corals that can be monocentric, but its calices grow larger (Ø > 50 mm) and may eventually form secondary stomata by intracalicular, circumoral budding. Other extinct agariciid genera predominantly consist of encrusting and massive, polystomatous species (Wells 1954, Veron 2000. Support from molecular analyses (Stefani et al. in prep) would be needed to justify the position of the new species in a separate genus instead of Leptoseris. In that case, it could be appropriate to classify it with Trochoseris Milne Edwards & Haime, 1849, an extinct genus (Mid Cretaceous -Oligocene) consisting of corals that are solitary, attached and turbinate or trochoid (Wells 1956). However, molecular analyses cannot support such a transition because no live material is available of this genus.
The corals have a basic growth form like that of Cladopsammia gracilis (Milne Edwards and Haime 1848) (Dendrophylliidae), i.e., by extramural budding from the basal coenosteum (see Cairns 1991). They are not distinctly reptoid as described for some Rhizopsammia species of the same family (Cairns and Zibrowius 1997) because there are no clear basal costate stolons involved. Extramural budding is also shown by fossils of the extinct genus Brachyphyllia Reuss, 1854 (Agariciidae), which has thick, low plocoid corallites (Wells 1956) instead of the thin cup-shaped corallites shown by the new species. Compared to the extant solitary agariciid deep-water species Dactylotrochus cervicornis (see Wells 1954, Cairns 1999, L. troglodyta differs by the capacity to become colonial and by having a circular corallum outline instead of a periphery with thecal extensions. Only a few specimens of D. cervicornis are known to show an 'aberrant' tendency for coloniality ).

Evolution of symbiosis with zooxanthellae
Leptoseris troglodyta is the first extant shallow-water agariciid known known to be reef-dwelling and azooxanthellate. The extinct agariciid genera, Trochoseris Milne Haime, 1849 andVaughanoseris Wells, 1934, also consists of monocentric species; the first being attached and turbinate or trochoid, and the second being free-living and discoid (Wells 1956). According to Kitahara et al. (2012) they could have been azooxanthellate because the extant Dactylotrochus cervicornis, which shows a basal position in the phylogeny reconstruction of the Agariciidae, is considered both azooxanthellate and monocentric. However, the phylogenetic positions of the two extinct taxa are unknown and although they share a (supposedly plesiosmorph) solitary growth form with Dactylotrochus, this does not necessarily imply that they are phylogenetically closely related to each other. For example, several phylogenetic lineages within the reef-dwelling mushroom coral family Fungiidae show an evolution from monocentric to polycentric zooxanthellate corals, implying that not all monocentric fungiid species are directly related to each other (Hoeksema 1989, 1991b, Gittenberger et al. 2011. Coincidently, the Fungiidae show many examples of solitary free-living species that resemble the extinct Vaughanoseris. So, regarding their lifestyle, Vaughanoseris species could be reef-dwelling and zooxanthellate as well. Moreover, the attached, monocentric dendrophylliid Balanophyllia europaea (Risso, 1826) is an example of a single zooxanthellate species (Schuhmacher and Zibrowius 1985) among a majority of congeners without zooaxanthellae . Its growth form resembles that of Trochoseris, for which the absence of zooxanthellae may therefore perhaps be less certain. Nevertheless, a solitary growth form, an old fossil record, and a possible ancestral position in the phylogeny may not be sufficient to predict whether an extinct coral taxon may have been azooxanthellate. Its habitat (especially depth) may be more indicative, especially if species were shallow reef-dwellers.
The deep-water species Dactylotrochus cervicornis and the cave-dweller Leptoseris troglodyta posses parallel ridges on the sides of their septa, which are called menianes (Kitahara et al. 2010. These are probably the same structures (compare Figure  5, Kitahara et al. 2012 figure 2, andKahng et al. 2012 figure 9) that help zooxanthellate Leptoseris corals to absorb sunlight more efficiently at greater depths . The bathymetric records of D. cervicornis  partly coincide with depth ranges of zooxanthellate Leptoseris species (see Dinesen et al. 2012). Depth may therefore not always be indicative for the absence of zooxanthellae.
All live specimens of L. troglodyta, which were observed in their habitat ( Figures  1-2), were clearly azooxanthellate by lacking a brown colour, like completely bleached corals (Hoeksema 1991a, Hoeksema andMatthews 2011). In the case of dredged deep sea corals in collections, which are either dry or being preserved in ethanol, pigments of any present zooaxanthellae might have dissolved, which makes it difficult to see whether they were present. If D. cervicornis corals lack zooxanthellae, like those of Leptoseris troglodyta, the presence of menianes at least suggests that their ancestors might have been zooxanthellate and that the loss of zooxanthellae may be an evolutionary novelty related to life in deep water or in caves. Consequently, ancestral agariciids, along with Trochoseris and Vaughanoseris, were perhaps also zooxanthellate like many modern monocentric zooxanthellate reef corals in illuminated habitats. On the other hand, a preceding presence of menianes in agariciid corals may also have facilitated the development of symbiosis with zooxanthellae, which implies that early agariciids may have been azooxanthellate as suggested by Kitahara et al. (2012), presenting a "chicken or the egg" causality dilemma.
The deep-water species Dactylotrochus cervicornis and the cave-dweller Leptoseris troglodyta both live in dark environments. The latter has been observed to be azooxanthellate in caves at various localities. D. cervicornis specimens may have to be examined for the absence of zooxanthellae to be sure that this species is always azooxanthellate. If so, its menianes have no use in connection to light absorption, like in L. troglodyta. L. troglodyta shows that the evolutionary relation between scleractinian reef corals and their algal symbionts is not fixed and that it may be difficult to deduct such a relation based on coral growth forms and their possible position in phylogeny reconstructions, especially if no molecular data can be obtained, as is the case for fossil corals. According to a recent molecular study, coloniality may have become lost at least six times and symbiosis with zooxanthellae at least three times in the phylogeny of the Scleractinia (Barbeitos et al. 2010). However, these numbers are based on a subset of species and may have to be revised if additional species are involved (see Gittenberger et al. 2011).

Eco-morphological considerations
Leptoseris troglodyta sp. n. is the first reef-dwelling agariciid coral without zooxanthellae. As a small cave-dwelling species it can live without sunlight. It has not been observed to co-occur with any zooxanthellate scleractinians (only azooxanthellate species), although in small and poorly illuminated cavities a variety of zooxanthellate scleractinians can be discerned (Dinesen 1983). Other Leptoseris species are zooxanthellate and most of them are able to live on deep, poorly illuminated reefs or rocky substrates. Therefore, from an evolutionary perspective, the new species may have lost the capacity to live in symbiosis with zooxanthellae and it may have obtained a smaller corallum size (dwarfism). By their small and thin corolla the corals have little weight. Consequently, with their wide basal plate they may not easily break off from their substrate, which consist of porous limestone cave ceilings where settlement space is limited. Without zooxanthellae, they may not easily reach large sizes, whatsoever. Owing to its modular growth form a L. troglodyta coral may risk losing a few expendable calices while other Leptoseris corals may harm and lose their entire corallum in unfavourable conditions. Other examples of zooxanthellae loss in relation to a cavernicolous lifestyle are so far unknown among scleractinian families. However, among shallow-water brachycnemic zoanthids an undescribed cave-dwelling Palythoa species has been recorded that also lacks zooxanthellae, while its congeners are known to be zooxanthellate (Reimer 2010).
the Scleractinia Systematics Working Group (SSWG) for their suggestions given during discussions. I thank Ms Zarinah Waheed and two reviewers, Dr Zena Dinesen and Dr Marcelo Kitahara, for constructive remarks on early drafts of the manuscript.