A centuries old problem in nephtheid taxonomy approached using DNA data (Coelenterata: Alcyonacea)

The current centuries old classi ﬁ cation of the family Nephtheidae is still mostly based on colony morphology. In this family the Indo-Pacifc genera Litophyton , Nephthea , Dendronephthya and Stereonephthya, and the Atlantic genus Neospongodes form a complex mix of closely related, poorly described species which cannot be recognized using only colony morphology. Attempts with the more modern approach of comparing skeleton composition (sclerites) have been only partly successful because of the extreme variation of sclerite forms present in these genera. The genus Chromonephthea Van Ofwegen, 2005, introduced for several species previously assigned to Dendronephthya , Nephthea and Stereonephthya , was established with sclerite morphology, but the true generic status of the majority of the nominal species of these genera remained unresolved. In an attempt to clarify the phylogenetic relationships between Litophyton , Nephthea , Stereonephthya and Chromonephthea four-teen specimens, unidenti ﬁ ed but certainly belonging to these genera, have been used in molecular analyses. All analyses supported two clades, which could be related to the shape of the sclerites present in the polyp stalks. One clade contained the specimens with characters for Stereonephthya along with the Chromonephthea specimen as a sister group. The other clade had a ‘true’ Nephthea and Litophyton together with the specimens that could not be placed in any particular genus using the old classi ﬁ cation criteria. The consequences of these results for nephtheid classi ﬁ cation are discussed. of


Introduction
After the Alcyoniidae Lamouroux, 1812, the family Nephtheidae Gray, 1862, is the most common soft coral family in tropical Indo-Pacifi c waters regarding numbers of genera and species. The often colourful colonies of nephtheid species can form large aggregates by asexual reproduction. Most species have treeor bush-like colonies and can be found from low tide to the abyss.
To date, Litophyton, Nephthea, Dendronephthya and Stereonephthya form a complex mix of closely related but poorly described species. Hence, around 1990, the fi rst author started a morphological revision of the genera Litophyton, Nephthea, and Stereonephthya, also including a number of species assigned to Dendronephthya and the monotypic Atlantic genus Neospongodes. A fi rst outcome of this study was the description of the genus Chromonephthea Van Ofwegen, 2005, for several species previously assigned to Dendronephthya, Nephthea and Stereonephthya.
In an attempt to clarify the phylogenetic relationships between some species that are actually classifi ed with Litophyton, Nephthea, Stereonephthya, and Chromonephthea, a few molecular markers previously tested with octocorals were used. By doing so the status of the morphological characters nowadays employed in nephtheids can be compared with the resulting phylogeny. Forskål (1775) described the fi rst Indo-Pacifi c nephtheid, Litophyton arboreum, from an unknown locality in the Red Sea. According to modern standards his description was inadequate. The material on which he based his description is lost.

Short history of the genera Litophyton, Nephthea, Stereonephthya, Dendronephthya and Neospongodes
Andouin (1828), referring to plates published by Savigny (1817), erected two additional nephtheid genera from the Red Sea, viz. Ammothea for A. virescens, and Nephthea, for N. chabrolii. As far as known, the material on which these two taxa were based has been also lost.
After this a period followed in which numerous nominal genera and subgenera were introduced or synonymized. Many new species were described, and others were moved from genus to genus. Kükenthal (1903Kükenthal ( , 1905) made a revision of the nephtheids in which he abandoned Ammothea and described three more genera, Neospongodes Kükenthal, 1903, Dendronephthya Kükenthal, 1905, and Stereonephthya Kükenthal, 1905, all of which without assigning a type species. He presented the following classifi cation, which is still used by most authors (for details of terminology, see Bayer et al. (1983)):  Kükenthal, 1905, as the type species of Stereonephthya. He characterized Stereonephthya as follows: (1) supporting bundles of the polyps well developed, each with a strong projecting tip, (2) ventral side of the polyp stalk with small tiny rods, (3) polyp heads usually hanging down from their stalks, like a campanulate fl ower, (4) dorsal points strong, ventral points poorly supported or lacking.
Finally, Verseveldt (1983) re-examined Neospongodes atlantica and N. bahiensis, the two species Kükenthal (1903) referred to Neospongodes. Verseveldt designated a lectotype for N. atlantica and synonymized N. bahiensis with N. atlantica. He changed the diagnosis of Neospongodes to "Tree-like, very fl abby nephtheids, with a slender stalk and stem. Polyps isolated and scattered on thin branches and twigs, with supporting bundles".

The modern classifi cation
Re-examination of most of the type specimens of the species described in Litophyton, Nephthea and Stereonephthya showed that the characters used by Kükenthal (1903Kükenthal ( , 1905) may be problematic: 1) The presence of a supporting bundle is not as straightforward a character as suggested by Kü kenthal (1905). In the re-examined holotype of Litophyton acutifolium Kükenthal, 1913, most polyps are without a trace of such a bundle, as diagnosed for Litophyton, but a few polyps clearly show one. Moreover, specimens identifi ed by Verseveldt (1966) as L. arboreum Forskål, 1775, have a supporting bundle, though Verseveldt reported it as missing.
Re-examination of the polyps of this species brought many small, irregularly arranged rods to light, but quite a few polyps show larger sclerites at the dorsal side, which in our opinion is a clear supporting bundle. 2) All nephtheids have an accumulation of polyps on the end of their branches. Species of Dendronephthya with polyps united in small bundles clearly differ from other nephtheids with respect to this character, but the presence or absence of catkins seems to rely on the amount of contraction after collection. Species with small polyps with less sclerites simply contract more easily to form catkins. 3) Polyps in bundles or single. This character is very diffi cult to check in Stereonephthya and species of the Divaricata group of Dendronephthya, as both have very crowded polyps at the end of the branches. Using the presence of small rods in the polyp stalk, as suggested by Utinomi (1954), the type species of Dendronephthya, D. savignyi (Ehrenberg, 1834) falls within Stereonephthya. 4) The presence of small rods in the polyp stalk in species of Stereonephthya. Examination of Dendronephthya species of both the Glomerata and the Umbellata group showed that polyps never have small rods in the polyp stalk, which are present in several species of the Divaricata group. Moreover, many species assigned to Nephthea also show small rods in the polyp stalk. 5) According to Verseveldt (1983), the differences between Neospongodes and Stereonephthya are minimal, in that the latter has rigid colonies and Neospongodes fl abby ones, which cannot be used to differentiate between genera convincingly. 6) Kükenthal (1903Kükenthal ( , 1905 completely ignored the sclerite characters of other parts of the colony. Using those the fi rst author recognized Chromonephthea Van Ofwegen, 2005.

Short descriptions of the material used
The material used for DNA analyses is briefl y characterized. Because of the above mentioned problems with the classifi cation and the still ongoing morphological revision, the species of Litophyton, Nephthea, and Stereonephthya could not be identifi ed. However, we do mention 'true' Litophyton and Nephthea following Kükenthal's classifi cation. The fl abby colony is 20 cm long with several stems arising from a common base (Fig. 1a). The polyps are grouped at the end of the branches, a confi guration described in literature as slender catkins. Its sclerites are typical for a species of Litophyton. The polyps and end branches have very few sclerites (Fig. 2a). The irregularly arranged polyp sclerites are small rods, which are up to 0.07 mm long (Fig.  3a). The surface layer of the base of the colony has radiates and derivatives of these ( Fig. 3b, d), and unilaterally spinose spindles with long spines (Fig.  3c). The interior of the base of the colony has spindles with sparse simple tubercles ( Fig. 3e- The colony used is 10 cm long (Fig. 1b). The polyps are grouped at the end of the branches, forming conical catkins. Its sclerites are typical for a species of Nephthea. The polyps have a supporting bundle and eight irregular points (Fig. 2b). The supporting bundle spindles have their distal end unilaterally spinose (Fig. 4c). The spindles of the ventral points are rod-like ( Fig. 4a, left), the dorsal points have unilaterally spinose spindles (Fig. 4a). The tentacles have fl attened rods with a scalloped edge (Fig. 4b).
Similarly shaped rods are also present at the ventral side of the polyp stalk (Fig. 2c). The surface layer of the base of the colony has radiates and derivatives of these ( Fig. 4d), and unilaterally spinose spindles with long spines (Fig. 4e). The interior of the base      ; e-f, spindles and unilaterally spinose spindles, surface layer base of colony; g, spindles, interior base of colony. Scale at c 0.01 mm, also applies to b; scales at e and g 0.10 mm; that at e also applies to a, d, that at g also to f. Fig. 8. RMNH Coel. 34811; a, point spindles, b, rods from ventral points and polyp stalk; c, tentacular rods; d, spindle of supporting bundle (part); e, spindles and unilaterally spinose spindles, surface layer base of colony; f-g, spindles, interior base of colony. Scale at c 0.01 mm, also applies to b; scales at e and f 0.10 mm; that at e also applies to a, d, g. Fig. 9. RMNH Coel. 34812; a, point spindles, b, rods from tentacles, ventral points and polyp stalk; c, spindle of supporting bundle; d-e, spindles and unilaterally spinose spindles, surface layer base of colony; f-g, spindles, interior base of colony. Scale at b 0.01 mm; scales at d and e 0.10 mm; that at d also applies to a, f; that at e also to c, g.
of the colony has branched spindles with sparse simple tubercles (Fig. 4f).
3) RMNH Coel. 34808, KOR.09, Palau, "Toagel Mlungi Channel", channel in western barrier reef off Babeldaob, 07°32'33.0"N 134°28'06.6"E; depth -25 m, coll. L.P. van Ofwegen,18.v.2005. The colony is 6 cm long (Fig. 1c). The polyps are grouped at the end of the branches, forming irregularly shaped catkins. The polyps have eight points and a slightly projecting supporting bundle (Fig.  2d). The point spindles (Fig. 5a) and supporting bundle spindles (Fig. 5c) are bigger than in the previous RMNH Coel. 32055, but basically the same types occur. The species differs from the foregoing in having more rods on the ventral side of the polyp stalk, which hardly differ from those found in the tentacles (Fig. 5b). The surface layer of the base of the colony has spindles (Fig. 5e), and weakly developed unilaterally spinose spindles (Fig. 5d), many of them with side branches. The interior of the base of the colony has spindles with sparse simple tubercles, only a few smaller ones are depicted (Fig. 5f).
They can be as long as 1.7 mm, and are less tuberculate than the surface spindles. The colony is 5 cm long (Fig. 1d). Like the previous two species the polyps are grouped at the end of the branches, forming irregularly shaped catkins. Also the polyp arrangement is much the same but differs in the ventral points and most of the polyp stalks having many rods (Figs. 2e-f, 6a-c). The surface layer of the base of the colony has spindles and unilaterally spinose spindles (Fig. 6d-e), many of them with side branches. The interior of the base of the colony has spindles with sparse simple tubercles (Fig. 6f), also many with side branches. The colony is 11 cm long, with several stems arising from a common base (Fig. 1h). The catkin shape and polyp sclerite arrangement (Fig. 2h) is very much the same as in the previous species. Also the polyp sclerites are similar ( Fig. 8a-d). However, the sclerites of the surface layer of the base of the colony are only slightly unilaterally spinose (Fig. 8e), and hardly any sclerites of the base have side branches ( Fig. 8f-g The colony is 4 cm long (Fig. 1f). The polyps appear very spiny due to strong supporting bundles (Fig. 9c), and an arrangement in catkins is not recognizable. The polyp armature is also different from the above described species, with point spindles slightly protruding beyond the polyp head (Fig. 2i). These spindles are not unilaterally spinose but have a spiny distal end (Fig. 9a). Tiny rods (Fig. 9b) are not only present in the tentacles, ventral points and ventral side of the supporting bundle, but also in between the points. The surface layer of the base has spindles and unilaterally spinose spindles (Fig.  9d-e), some of which are very large (Fig. 9e). The interior has spindles with sparse tuberculation (Fig.  9f-g The colony is 8 cm long, but the base of the stalk is probably lost (Fig. 1g). Like in the previous species, the supporting bundles are strong (Fig. 10e), and catkins are not clearly discernable. In this species small rods (Fig. 10b-d) are present all over the polyps (Fig. 2j), and the point spindles are unilaterally spinose (Fig. 10a). The surface layer of the base of the colony has spindles and unilaterally spinose spindles ( Fig. 10f-g), the latter poorly developed. The interior has spindles with sparse tuberculation (Fig. 10h). The sclerites at the base of the colony look more like the types that are normally present higher up in the stalk of nephtheids. Hence we do not exclude the possibility that the lower part of the colony is lost. The colony is 6 cm long with a colony shape very much alike that of RMNH Coel. 34812, but with more slender lobes; catkins are not recognizable, and the stalk is covered by sponge tissue (Fig. 11c). Also the polyp armature (Figs. 12a, 13b-d) and shape of the point spindles (Fig. 13b) are similar to those of RMNH Coel. 34812. It differs by having the point sclerites protruding more beyond the polyp body, having less small rods in the polyp (Fig.  12a), and these rods being of a different shape (Fig.  13a). The surface layer of the base of the colony has radiates and derivatives of these, spindles and poorly developed unilaterally spinose spindles (Fig.  13e). The interior of the base of the colony has spindles with sparse tuberculation (Fig. 13f) The colony is 10 cm high, yellow with white polyps, and a real stalk is lacking (Fig. 11b). The colony shape is typical for a species of Stereonephthya with the polyps making an acute angle with their polyp stalk and supporting bundles projecting considerably beyond the polyps (Figs. 12b, 14d). The points have very big spindles, some of which also project beyond the polyp with the projecting part often widened and less spiny (Fig. 14b). The polyp stalk and tentacle rods (Fig. 14a, c) are similar to those of the previous species. The surface layer of the base of the colony has radiates, derivatives of these, spindles, and poorly developed unilaterally spinose spindles (Fig. 14e-f). The interior of the base of the colony has very few spindles, and these have sparse tuberculation (not depicted). The colony is 9 cm long, white with orange polyps, with typical Stereonephthya colony shape (Fig.  11a). The polyp armature (Fig. 15a-d) is like the previous species but the point sclerites are much smaller (Fig. 15b) and they do not project beyond the polyp (Fig. 12d). The sclerites of the surface layer of the base of the colony are more slender than in the previous species (Fig. 15e-f). The interior of the base of the colony has spindles with sparse tuberculation (Fig. 15g) (2005), and here the holotype (Fig. 11d) and polyp armature (Fig. 12e) are shown for comparison. The genus differs from Nephthea and Litophyton by being azooxanthellate. It differs from Stereonephthya and Neospongodes by lacking rods in the polyp stalk, and from Dendronephthya by having the polyps standing single instead of in small bundles. Fig. 10. RMNH Coel. 34813; a, point spindles; b, rods from polyp stalk; c-d, tentacular rods; e, spindle of supporting bundle (part); f-g, spindles and unilaterally spinose spindles, surface layer base of colony; h, spindles, interior base of colony. Scale at b 0.01 mm, also applies to c; scales at f and g 0.10 mm; that at f also applies to a, d-e; that at g also to h.    ; e-f, radiates, spindles and unilaterally spinose spindles, surface layer base of colony; Scale at c 0.01 mm, also applies to a; scales at e and f 0.10 mm; that at e also applies to b, d. ; e-f, radiates, spindles and unilaterally spinose spindles, surface layer base of colony; g, spindles, interior layer base of colony. Scale at c 0.01 mm, also applies to a; scales at e and f 0.10 mm; that at e also applies to b, d, g.

Methods
A small branch of each coral was preserved in 96% ethanol for molecular analysis. Around 0,25 cm 3 of tissue was cut into small pieces for surface enlargement. Genomic DNA was extracted by use of a DN-Easy ® Tissue Kit (Qiagen) following the manufacturer's protocol for animal tissues. Elution volume was 50-100 μl and the average DNA concentration was around 1250 ng/μl. Best PCR results were obtained using 1:10 dilutions and for most extracts, PCR's failed without the addition of Q-solution (Qiagen). Co-extraction of zooxanthellae was unavoidable and shown when we amplifi ed the ITS-1 and ITS-2 regions. Initial attempts to sequence mitochondrial marker MSH with primers ND42599F 5'-GCCATTATGGT-TAACTATTAC-3' and Mut-3458R 5'-TSGAG-CAAAAGCCACTCC-3' Hoover, 2001, 2002) failed for most taxa. Therefore we designed internal primers MSH-Neph-F 5'-TATGAACTTT-GGCATGAGCC-3' and MSH-Neph-R 5'-TGCCC AAATTACTATTTCTCTAATACG-3' specifi c for these nephtheids. PCR-reactions were done using a standard Taq DNA polymerase kit from Qiagen. Reaction volume always was 25 μl and PCR conditions were 0.2 mM DNTPs, 0.4 μM of each primer, 5 μl of Q-solution, 5 units of Taq DNA polymerase (Qiagen) and no additional MgCl 2 (PCR-buffer always contains 1.5 mM MgCl 2 ). Thermocycling conditions were 3 min. at 94ºC for initial denaturation followed by 40 cycles (15 sec. at 94ºC, 1 min. 30 sec. at 50ºC and 1 min. at 72ºC) and fi nal extension 5 min. at 72ºC. PCR products were purifi ed using Nucleospin ® kit (Macherey-Nagel) and cycle sequenced in both directions with an ABI Dye Terminator Cycle Sequencing Kit (PE Biosystems). The sequenced products were purifi ed using G-50 Fine Sephadex ® columns and run on an ABI 377 automated sequencer. Sequences were checked and edited with Sequencer 4.2 (Genecodes inc.).

Sequence variation
The sequences we used ranged from 860 bp (RMNH Coel. 34814) to 977 bp (RMNH Coel. 32053) in length. All sequences included the 157 bp of the 5.8S gene, the 3' terminus of 18S ranged from 114 bp (most sequences) to 117 bp (RMNH Coel. 32050), but was mostly 114 bp long. The 5' end of 28S was Table 1. Variation in length and base composition of ribosomal internal transcribed spacer regions (ITS-1, ITS-2) among species of nephtheids. Alcyonium digitatum is included as outgroup. bp = base pairs, CG = cytosine and guanine. The long sequence length of RMNH Coel. 32053 with 977 bp is caused by its long ITS regions. In the ITS-2 region at position 547 there is a fi ve times repeat of "GGCCCTTT", and at position 704 a three times repeat of "TAAAGCTCTT", which mostly explains for the extraordinary length. In the ITS-1 region such obvious repeats were not recognized. RMNH Coel. 32058-59 and RMNH Coel. 34814-16 seem to have a shorter ITS2 region (on average 186.20) than the other nephtheids (on average 214.33).
The G+C content of ITS1 and ITS2 was calculated in GeneDoc (Table 1), the values for ITS-1 are comparable with those given for Alcyonium species by Mc-Fadden et al. (2001), the values for the ITS-2 region are on average more than 10% higher. No pattern could be recognized within the nephtheids.
Nucleotide diversity was estimated in PAUP using the Kimura two-parameter method (Table 2). Genetic distances of the combined partial sequence of the 5' end of 18S rRNA gene, the complete sequence of ITS-1, 5.8S rRNA gene and ITS-2, and a partial sequence 3' end of 28S rRNA are presented in Table 2. The distance between RMNH Coel. 32058 and RMNH Coel. 34815 is 0% and that between these two and RMNH Coel. 34814, 16 and RMNH Coel. 32059 is also very small, 0.0023-0.0094.
Sequences were aligned using ClustalW (Thompson et al. 1994) with parameter GOP (Gap Opening Penalty) set to 2 and the GEP (Gap Extension Penalty) to 1. Both McFadden et al. (2001) and Forsman et al. (2006) proved that this is the best alignment setting for Alcyonium (soft corals) and Porites/Siderastrea (stony corals).
Genbank accession numbers are presented in Table 3.

Phylogenetic analyses
We used the permutation test in PAUP (version 4.0b10; Swofford, 2002) for a phylogenetic signal, the P value was 0.01.   The fi nal alignment included 1029 nucleotide positions: 685 are constant, 123 are variable, and 221 are parsimony-informative.
PAUP was used for maximum parsimony, neighbour-joining, and the Unweighted Pair Group Method with Arithmetic mean (UPGMA). For maximum parsimony we did bootstrap analyses (100 and 1000 replicates) using the heuristic-search option with simple addition of sequences and Tree-bisection-reconnection (TBR) branch swapping. The steepest descent option was in effect. Branches having maximum length zero were collapsed to yield polytomies. Multrees' option was in effect.
For neighbour-joining we used bootstrap analyses (1000 replicates) with consensus tree option set to 50% majority-rule, for the rest we used default settings.
With MacClade (version 4.08) we examined the base pair positions used in the neighbour-joining tree analysis. As can be seen in Fig. 16 this mostly concerned positions in the poorly aligned ITS regions. Therefore, we also did the above analyses after removing the less aligned parts of the ITS regions.
Alcyonium digitatum was used to root all trees.

Results
The different number of replicates used in the bootstrap analysis of maximum parsimony resulted in only marginal differences. The different methods yielded similar results. In all trees two clades are supported. In all but the partial ITS parsimony analysis clade I is formed by Chromonephthea franseni together with RMNH Coel. 32058-59 and RMNH Coel. 34814-16 and clade II includes all other specimens. Only in the partial ITS parsimony anal-  ysis Chromonephthea franseni clustered in clade II.
Here we only present the results of the UPGMA analysis (Fig. 17). Because of the poor alignment of the ITS regions, which are mostly used in the analyses (Fig. 16), we looked at overall sequence similarity rather than evolutionary relationships. However, the maximum parsimony and neighbour-joining bootstrap analyses gave similar results.

Discussion
The two distinct clades that show up in the analyses were compared with the morphological data. The stalk sclerites, point spindles, and colony form did not yield any clear correlation with the phylogeny found. However, specimens in Clade I, with the exception of the Chromonephthea franseni specimen, showed the characters for Stereonephthya as reported by Utinomy: 1) supporting bundles of the polyps are well developed with strong projecting tip; 2) ventral side of the polyp stalk has small tiny rods; 3) polyp heads usually hang down on their stalks, like a campanulate fl ower; 4) dorsal points strong, ventral points poorly supported or lacking. RMNH Coel. 34814 has both 'campanulatelike' polyps and ones like depicted in fi gure 12a. The only consistent sclerite character found for clade I is the shape of the rods in the polyp stalks, in this clade they have tiny projections, and are longer (Figs. 13a, 14a, 15a) than those in clade II.
The polytomy of RMNH Coel. 34815-16 and RMNH Coel. 32058 in most of the analyses (Fig. 16) confi rms the identifi cation of these three specimens based on sclerite characters, which showed them to belong to a single species.
The grouping of Chromonephthea franseni as the sister group next to the Stereonephthya species probably occurred because only one sequence of Chromonephthea was included.
Clade II is a mixture of specimens which, on the basis of the characters used in the classifi cation of Kükenthal, could be placed in Nephthea, Litophyton or Stereonephthya. The Litophyton-like specimen (polyps in catkins, lacking supporting bundle) and the Nephthea-like specimen (polyps in catkins, supporting bundle) group with Stereonephthya-like specimens (projecting supporting bundle, catkins not obvious, tiny rods in the polyp stalk). All species in Clade II lack the long polyp stalk rods with tiny projections.
Obviously, the sclerite character found to support the two different clades is extremely small for separat-ing different genera in nephtheids, and therefore it is not surprising that previous researchers never mentioned it as a character. However, despite the small difference in shape, the polyp stalk rods could have a different origin. In all species included in Clade II the polyp stalk rods seem to be similar to those found in the tentacles and ventral points, while in Clade I they are situated somewhat further apart from the polyp head and are clearly different from the tentacle and ventral point rods. They could represent extremely reduced polyp stalk spindles. Mcfadden et al. (2006), in their phylogenetic analysis of octocoral genera, used the mitochondrial markers ND2 and MSH combined. They also included specimens of Litophyton, Nephthea and Stereonephthya in their analyses. Like we found with MSH, they also found identical sequences for Litophyton and Nephthea. In their analyses they always found a polytomy of Litophyton/Nephthea, Stereonephthya and a specimen identifi ed as belonging to the Atlantic genus Neospongodes. As the 'Neospongodes' specimen was collected in Darwin harbour the fi rst author contacted one of the authors of McFadden et al., viz. Phil Alderslade (Museum and Art Gallery of the Northern Territory, Darwin, Australia), who informed us that the 'Neospongodes' specimen actually belonged to Chromonephthea. The Litophyton, Nephthea and Stereonephthya specimens used were identifi ed with the same criteria we applied. The fi ndings of McFadden et al. do not contradict ours, but apparently the 18S-ITS1-5.8S-ITS2-28S marker we used gives more resolution than the combined ND2 and MSH markers.
The results of the present molecular study have consequences for the nephtheid classifi cation. In the morphological revisions under progress Nephthea should be synonymized with Litophyton. All species previously assigned to Stereonephthya and Nephthea should be checked with regards to the shape of the polyp stalk rods.
Moreover, Nephthea savignyi Ehrenberg, 1834, the type species of Dendronephthya, has the sclerite and colony form characters of clade II species and should therefore be transferred to Litophyton, making Dendronephthya a junior synonym of Litophyton. Finally, in the Atlantic genus Neospongodes the polyp stalk sclerites are similar to those found in clade I (Indo-Pacifi c Stereonephthya) and, according to the code of nomenclature, this name has preference above Stereonephthya. Unfortunately, no recently collected material of Neospongodes species was available for the present DNA study.
For the moment, pending the completion of the morphological revisions in progress, we retain to the current classifi cation.