A key to the genera and species of the transversely-dividing Flabellidae (Anthozoa, Scleractinia, Flabellidae), with a guide to the literature, and the description of two new species

Abstract The transversely-dividing flabellids consist of five genera (Truncatoflabellum, Placotrochides, Blastotrochus, Placotrochus, and Falcatoflabellum) and 45 species. A dichotomous key is provided for these five genera as well as the species of the genus Truncatoflabellum and Placotrochides, the other three genera being monotypic. A tabular key is also provided for the 38 species of Truncatoflabellum. Two new combinations are suggested (Truncatoflabellum gambierense and Truncatoflabellum sphenodeum) and two new species are described (Truncatoflabellum duncani and Truncatoflabellum mozambiquensis). All but one species are illustrated and accompanied by their known distribution and a guide to the pertinent literature for the species. New records of 19 of the 45 species are listed. The transversely-dividing flabellids range from the Middle Eocene to the Recent at depths of 2–3010 m, and constitute 60% of the 65 known extant species of transversely-dividing Scleractinia.


Introduction
Confronted with a large collection of Truncatoflabellum during a recent (2014) trip to Taiwan, it became apparent that the literature on the species of this genus was scattered and not well organized. Although there were some keys to the species, they were regional in nature, i.e., Philippine region (Cairns 1989b), southwest Indian Ocean (Cairns and Keller 1993), North Pacific (Cairns 1994), and Western Australia (Cairns grove (1938) reported eight specimens as Flabellum rubrum from the Pleistocene of Talaud, Celebes, one of which is T. aculeatum, four of which are unidentifiable to species, and three are Trochocyathus. Although not illustrated by Umbgrove (1938), these specimens are also deposited at the RGM (35461). Umbgrove (1950) also reported two Truncatoflabellum species from the Lower Pleistocene of the Putjangan Beds of Java, part of one of which has been re-identified as T. carinatum (see Cairns, 1989b). Those specimens were deposited at the Institute of Mines at Delft in 1989. Various species of Truncatoflabellum from the Pliocene of Taiwan and Plio-Pleistocene of the Ryukyu Islands were reported by Yabe and Eguchi (1942a, b) under the rubric of F. rubrum. Most of these specimens, deposited at the TIUS, were examined by the author and reidentified in Cairns (1989b). Yabe and Eguchi (1941) also reported one fossil Truncatoflabellum (=T. spheniscus) from the "Neogene" of Java. Squires (1958: pl. 12, figs 6-7) illustrated two Truncatoflabellum from the Altonian (Lower Miocene) of New Zealand as Flabellum rubrum rubrum, but these are certainly not F. rubrum and have not been subsequently re-identified. Most specimens from that paper are deposited at the AUC. Hayward reported F. sphenodeum (=T. sphenodeum) from the Lower Miocene of North Auckland, New Zealand; these specimens are also deposited at the AUC. Wells (1984) reported two species from the Late Pleistocene deposits of Kere River, Santo, Vanuatu: F. vanuatu (=T. vanuatu) and F. paripavoninum (=T. mortenseni); these specimens are deposited at the NMNH. Finally, Hu (1987) reported two truncate flabellids from the Maanshan Mudstone (Plio-Pleistocene) of Hengchun Peninsula, southern Taiwan: Flabellum transversale (=T. carinatum) and Flabellum elongatum (=Placotrochides scaphula); these specimens are deposited at the National Museum of Natural Science, Taichung, Taiwan, and seen by the author in 2014. Hu (1988) also reported Flabellum rubrum stokesii from the Tunghsiao and Lungkang Pleistocene formations of the Miaoli District, northern Taiwan, some of which are probably T. carinatum.
It should be noted that in a comprehensive phylogenetic analysis of the Scleractinia using the CO1 gene (Kitahara et al. 2010), in which 65 additional deep-water species were included to the data base, Truncatoflabellum was found to be polyphyletic and always ancestral to species within the genus Flabellum. Both genera have their earliest records in the Eocene.

Methods
This is not a taxonomic revision or a phylogenetic or morphometric analysis. It is a key to facilitate identification of a species-rich group, accompanied with a guide to the literature. The synonymies are not exhaustive, but include the original description and those papers that were found useful in identification of the species, especially those that contain useful illustrations, descriptions and/or extended synonymy. Furthermore, the key incorporates exclusively fossil species that occur in the respective genera. Since the key is intended to serve a practical purpose and include fossil species, molecular sequencing was not employed.
In an effort to discuss and illustrate morphologically similar species in adjacent text, and to facilitate their identification through keys, the text and illustrations are arranged in the order in which they occur in the key.
The key is based primarily on the morphology of the (free-living) anthocyathus stage of each species, the founding (attached) anthocaulus stage rarely being collected and usually of generic morphology. The shape of the anthocyathus contains the primary distinguishing set of characters for these genera, the shape most accurately defined by the thecal edge and face angles (Fig. 1). These two measurements geometrically define the GCD:LCD, and thus that index is not an independent one, but is presented in Table 2 because of its ease in visualization. The H:GCD is a general measure of the height of the corallum, but is dependent on the size (maturity) of the corallum, thus adult specimens are best measured for this characteristic. The maximum greater scar diameter (GSD), on the other hand, is fairly constant, being the same size for juvenile or large specimens; however, the ratio of GSD:GCD is dependent on the size of the corallum. In addition to shape criteria, the number of pairs of thecal edge spines seems to be relatively constant, some species having none, others one basal pair, others four or more pairs, and still others two or three pairs. The purpose of the thecal edge spines is unknown, however Tokuda et al. (2010) suggest that they function to stabilize "the life position" of the anthocyathus after transverse division. Several species have crests instead of spines. Other characters useful in differentiating species are: number and symmetry of the septa, nature of the upper outer edges of the septa as they meet the theca (e.g., notched, attenuate, abrupt), and corallum color. Geography, fossil occurrence, and even depth distribution may also be used as circumstantial characters.
Of the 45 species of truncate flabellids, 41 are represented in the NMNH collections, including types of 27 of those species. Of those four species not represented in the NMNH collections, photographs were obtained of three (T. inconstans, T. gippslandicum, and T. sphenodeum); only T. trapezoideum (known only from one specimen deposited in Moscow) was not re-examined and not illustrated herein. Whenever possible, five views of a typical anthocyathus of each species is presented in a vertical arrangement, top to bottom: lateral face, edge, basal scar, calice, and oblique calice.

Diagnosis.
Asexual reproduction by apical transverse division of corallum, resulting in distal anthocyathus and basal anthocaulus. Corallum usually laterally compressed and fan shaped, having one or more pairs of thecal edge spines or crests; some species compressed-cylindrical in shape but these always laterally spinose, whereas some fanshaped coralla lack spines and crests. Columella absent or represented by a fusion of the lower, axial edges of larger septa. Anthocaulus not stereome-reinforced. Discussion. The taxonomic history of this genus extends long before it was officially described, and is recounted and discussed by Cairns (1989b). To briefly re-iterate, even as early as 1848 Milne Edwards and Haime (1848) placed these species in a section (=subgenus) they called the "flabelline tronquees". Squires (1963: 10, 25) strongly felt that this group of species should be separated as a genus different from Flabellum but ultimately did not take an action, waiting for more biological justification. In Zibrowius' (1974) revision of the family Flabellidae, he placed the transversely-dividing Flabellum as one of three "groups" in the larger conventional genus Flabellum. Finally, in a paper about the various modes of asexual reproduction, Cairns (1989a) suggested that transverse division represented a key innovation that led to an adaptive advantage for living on soft substrates, justifying the naming of a new genus. But, it was not until later in that year that Cairns (1989b) proposed the name Truncatoflabellum. As of this paper, there are 38 known species in the genus, six of these known only as fossils (Table 1) (Alcock, 1894) dens (Alcock, 1902) incrustatum Cairns, 1989 =+irregulare sensu Gerth, 1921:402 (new synonymy) formosum Cairns, 1989 =T. sp. n. sensu Cairns, 1989:73 pusillum Cairns, 1989 carinatum Cairns, 1989 ?+variable alta Gerth, 1921, if so, name is altum angustum Cairns & Zibrowius, 1997 Central and eastern Pacific trapezoideum (Keller, 1981) truncum (Cairns, 1982) Vanuatu, Wallis andFutuna, New Caledonia martensii (Studer, 1878) =+paripavoninum sensu Wells, 1984 mortenseni Cairns & Zibrowius, 1997 vanuatu (Wells, 1984) vigintifarium Cairns, 1999 New Zealand and Kermadecs arcuatum Cairns, 1995phoenix Cairns, 1995 =T. sp. B sensu Cairns, 1994 Western Australia angiostomum (Folkeson, 1919) australiensis Cairns, 1998 veroni Cairns, 1998 macroeschara Cairns, 1998 Western Indian Ocean/S. Africa stabile (Marenzeller, 1904) =Truncatoflabellum sp. A sensu Cairns, 1994: 79 =?T. sp. Zibrowius & Gili, 1990inconstans (Marenzeller, 1904 gardineri Cairns in Cairns & Keller, 1993 zuluense Cairns in Cairns & Keller, 1993multispinosum Cairns in Cairns & Keller, 1993 mozambiquensis sp. n. South Australian exclusively fossil species +victoriae (Duncan, 1864) =?F. simplex Tenison-Woods, 1878 +gambierense (Duncan, 1864)  Remarks. This is the smallest of the Truncatoflabellum species, having a GCD rarely more than 5 mm, but capable of multiple apical regeneration ( Fig. 2A, top) resulting in coralla as long as 17.5 mm.  Remarks. The two syntypes of T. gippslandicum were reported by Bell (1981) from the NMV (P27064). No other records of this species are known, and the information presented in the key and comparative Table 2 is taken from the original description.

Distribution. Off Western Australia and Queensland, 18-201 m.
Remarks. T. macroeschara belongs to a group of three western Australian species that have very large coralla, often including some S7, the other two species being T. veroni and T. angiostomum. It differs from those two species as well as all others in the genus by having a very large scar diameter.   Remarks. In the original description, Duncan (1864) described the species as not having thecal edge spines, but in 1870 said that the coral has "often small spines nearer the calice than the pedicel." Indeed, some specimens of this distinctively-shaped species have spines (traditional Truncatoflabellum) and others do not (see New Records). Ordinarily, if a species of Truncatoflabellum bears thecal edge spines then all specimens of that species will bear spines. Thus, this variation in character is unusual and may be indicative of the early evolution in the genus when spination and transverse division were still experimental, as T. gambierense is one of those species that shows a crescentric transverse weakness in its corallum but the anthocyathus usually remains attached to the anthocaulus, possibly the ancestral condition for the species.    (Gerth, 1921). Holocene: Philippines; Indonesia; Ryukyu Islands, Japan, 30-315 m.
Remarks. According to the records of Squires (1958), this would be the oldest Truncatoflabellum, being reported from the Bortonian (Middle Eocene) of New Zealand.
The specimens reported by Hayward (1977) as "Flabellum" sphenodeum and "Flabellum" sp. A have much larger basal scars and shorter coralla than typical F. sphenodeum and are thus not included with this species. Specimens in the NMNH that may be the same are USNM 67932 and 67928, and may represent an undescribed species.
Anthocauli are rare, only four of the 262 (1.5%) specimens representing this juvenile stage. It is small, only about 4.1 mm in height with a circular attached pedicel 2 mm in diameter, and a distal calice 5-6 mm in greater diameter corresponding to the scar diameter of the anthocyathus. It has three cycles of septa.

Remarks.
As suggested by the key, T. mozambiquensis is most similar to T. martensii, but can be distinguished by its smaller basal scar, higher H:GCD ratio, rounded thecal edges, and tendency to have one (or occasionally two) pairs of thecal edge spines vs. three pairs for T. martensii (Table 2).

Truncatoflabellum vigintifarium
Remarks. The name spheniscus, Latin for small wedge, is treated as a noun in apposition and thus does not match gender with the genus.

Fig. 8B
Flabellum vanuatu Wells, 1984: 215, figs 4 (11-12), 5 (1). Truncatoflabellum vanuatu: Cairns 1989b, Table 6  similar in size to the specimen reported by Duncan. The GCD:LCD ratio is 1.5-2.1; the H:GCD = 0.95-1.05; and the GSD:GCD is about 0.27, with the scar reaching as long as 12 mm. Four or five pairs of prominent flattened thecal edge spines are present. The septa are quite regularly arranged in five cycles (S1-3>S4>S5), with one pair of S6 in each of the four end half-systems, resulting in 104 septa. The lower axial edges of the larger septa are only slightly sinuous, whereas the upper outer edges are gracefully attenuate, meeting the upper theca as low lamellae. The fossa is open, bordered by the axial edges of the wide S1-3. The anthocaulus is unknown.
Distribution. Late Oligocene to Middle Miocene, Victoria. Remarks. As suggested by the key, T. duncani is remarkably similar to T. multispinosum, but can be distinguished by its attenuated upper septal margins. It is also known only from the Oligocene to Miocene of Australia, whereas T. multispinosum is restricted to the Holocene and Late Pleistocene.
Etymology. Named in honor Peter M. Duncan, who first discovered specimens belonging to this species.
Remarks. Truncatoflabellum paripavoninum belongs to a group of six species that lack thecal edge spines and crests (see Key: couplets 28-32). Except for T. inconstans, known only from limited material from 23-130 m, these species have the greatest depth ranges of all the species in the genus often occurring deeper than 1000 m, suggesting that spines are less necessary for life at great depths. This begs the question of the function of the thecal edge spines. Even the relatively shallow species that have edge spines live at hundreds of meters of depth, far below the level at which surface turbulence would affect them. Thus the function of the thecal spines still remains unresolved.
Remarks. This is the deepest living Truncatoflabellum as well as the most geographically widespread.
Very rarely a pair of very small basal thecal spines may be present, but the species is considered to lack spines for the purpose of the key.
Remarks. The name corbicula, Latin for small basket, is treated as a noun in apposition and thus does not match gender with the genus.
Remarks. This species is known only from its original description.

Diagnosis. Like
Truncatoflabellum, but also producing asexual buds (anthoblasts) from thecal edges of anthocyathus. Thecal edges rounded, have a low edge angle, and bear one pair of basal edge spines. Discussion. The mode of asexual reproduction employed by Blastotrochus, described and illustrated by Cairns (1989a) as the anthoblast mode (also called bud shedding), differs slightly from transverse division of Truncatoflabellum by its potential to produce many more simultaneous clonemates from its thecal edges (instead of one at a time as with Truncatoflabellum), leading to a potentially exponential increase in clonemates instead of a gradual one. This was considered as a key innovation by Cairns (1989a), worthy of generic distinction from Truncatoflabellum. A second species was described in this genus, B. proliferus d'Archiardi, 1866 (Miocene, Italy), but was reassigned to Cladocora (see Pfister 1980). Blastotrochus thus remains a monophyletic genus and has rarely been collected. Distribution. As for the genus.

Diagnosis.
Asexual reproduction by apical transverse division of corallum, resulting in distal anthocyathus and basal anthocaulus. Corallum usually laterally compressed and subcylindrical, having a low edge angle; thecal edges rounded and do not bear spines or crests; calicular outline often asymmetrical. Columella absent of represented by a fusion of the lower, axial edges of the larger septa. Anthocaulus stereome-reinforced.
Distribution. Western and central Pacific, southwestern Indian Ocean, northern and southwestern Atlantic, 80-1628 m.

Diagnosis.
Asexual reproduction by apical transverse division of corallum, resulting in distal anthocyathus and basal anthocaulus. Corallum laterally compressed and fan shaped, having rounded thecal edges with one pair of basal thecal edge spines. Columella lamellar. Anthocaulus not stereome-reinforced.
Discussion. Seven species of Placotrochus were described from the Australian Eocene-Miocene by Duncan (1864), Dennant (1899Dennant ( , 1903Dennant ( , 1904, and Tenison-Woods (1878a), but these species are not transversely dividing and thus should be assigned to a different genus (Cairns in prep.). Placotrochus is a monotypic genus.

Distribution. As for genus.
Remarks. Known only from the type series of 21 specimens from the type-locality.