The material studied was collected on the tip of the Antarctic Peninsula, South Shetland Islands, near the South Orkney Islands, the South Sandwich Islands and the Scotia Sea (Fig. 1) on board the R/V Polarstern on the cruises ANT XIX/5 (LAMPOS, 3 April to 5 May 2002) and ANT XXIX/3 (22 January to 18 March 2013) sponsored by the Alfred Wegener Institut für Polar- und Meeresforschung (Bremerhaven) under the auspices of the Scientific Committee for Antarctic Research.

Octocoral colonies were collected using Agassiz trawls in the Bransfield Strait and South Shetland Islands area (Fig. 1). The octocorals were fixed on board in hexamethylenetetramine-buffered 10% formalin-seawater for morphological examination, or directly in absolute ethanol for further molecular studies.

Total genomic DNA was extracted from ethanol (EtOH)-preserved material using the E.Z.N.A. DNA kit (OmegaBiotech) following the manufacturer’s instructions. The mtMutS, ND2 and COI mitochondrial regions and the nuclear 28S were sequenced. The start of the mtMutS region was amplified using the primers ND42625F and MUT3458R (McFadden et al. 2006; Sánchez et al. 2003). ND2 was amplified using the primers 16S647F and ND21418R (McFadden et al. 2004). The COI region was amplified using the primers COII8068F and COIOCTR (McFadden et al. 2004; France and Hoover 2002). The 28S nuclear ribosomal gene (28S rDNA) was amplified using the primers 28S-Far and 28S-Rar (McFadden and van Ofwegen 2013). Each PCR used 1 U of MyTaq Red DNA Polymerase (Bioline), 10 μM of each primer and approximately 30 ng of genomic DNA, being brought to a final volume of 25 µL with molecular grade H₂O. MtMutS PCR was carried out using the following cycle profile: initial denaturation at 94°C for 2 min, 35 cycles of denaturation at 94°C for 15 s, annealing at 55°C for 15 s, extension at 72°C for 10 s and a final extension at 72°C for 5 min. The ND2, COI and 28S PCR used the same cycle profile, but the corresponding annealing temperatures were 51°C, 50°C and 58°C respectively. The PCR products were purified using ExoSAP-IT™ PCR Product Cleanup Reagent (ThermoFisher Scientific) following the manufacturer’s instructions before strong amplifications were sent to Macrogen Europe for sequencing in both directions. The purified products were electrophoresed on an ABI PRISM® 3730xl Genetic Analyzer.

All chromatograms were visualized and sequence pairs were matched and edited using Sequencher v4.0. A preliminary BLAST search identified our sequences close to the genera Kophobelemnon, Funiculina and Gyrophyllum, all them within the informally named Clade III. This relationship was also in an ML comparison based only on mtMutS (~400 pennatulacean sequences, not shown). Therefore, for the present study, we will only include sequences of Clade III to avoid repeating trees and discussions between clades already included in other recently published studies (e.g. Dolan et al. 2013, Kushida and Reimer 2019, Kushida et al. 2022, García-Cárdenas et al. 2020).

The set of new sequences and the homologous ones from GenBank belonging to related genera and species of sea pens within Clade III (Table 1) were aligned using MUSCLE (MEGA6, Tamura et al. 2013). After alignment, pairwise genetic distances based on the mtMutS marker and the Kimura 2-parameter (K2P) model of nucleotide substitution (Kimura 1980) were obtained to compare them with previous analyses at genus and family levels, following the comparisons of Pante et al. (2012), López-González (2020) and López-González and Drewery (2022).

In accordance with previously published molecular phylogenies on pennatuloids (e.g. Dolan et al. 2013, Kushida and Reimer 2019, García-Cárdenas et al. 2020, López-González and Drewery 2022), sequences of ellisellids from GenBank were selected as out-groups. Three data sets of sequences were analysed: mtMutS separately, the three mitochondrial genes mtMutS+ND2+COI, and the last one with the addition of the nuclear gene 28S. The matrix including only the mtMutS sequences had 689 bases. The concatenated matrix including only the three mitochondrial sequences had 2014 bases. The concatenated matrix for the mtMutS+ND2+COI+28S sequences had 2814 bases. All data matrices had 26 pennatuloid plus three ellisellid sequences.

The phylogenetic reconstructions were obtained by applying the BI and ML methods. The best nucleotide substitution model was selected using Modeltest (Darriba et al. 2012) according to the Akaike information criterion and hierarchical likelihood ratio test (hLRT) values (T92 for all the data matrices). ML was implemented in MEGA6 using the NNI heuristic method (nearest neighbour interchange) and 1000 bootstrap replications (Felsenstein 1985). The BI was carried out in the MrBayes v3.1.2 program (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003), using the model GTR+G (lset nst=6 rates=gamma) with 10⁷ generations, discarding 25% of the initial trees. The stationarity of the chains and the convergence of the two runs were monitored for each parameter by Tracer (v.1.7.1) (Rambaut et al. 2018) to ensure that the effective sample size of all parameters was larger than 200, as recommended.

Colonies were dissected under a Motic MSZ168 stereomicroscope, and fragments from different part of the colonies were prepared for corroboration of the presence of sclerites and sclerite preparation for light microscopy and scanning electron microscope observation. Fragments of different parts (tentacles, anthocodia, etc.) were prepared in clove oil to observe sclerite arrangement in a Leica DMLB light microscope with an OPTIKA C-P20CC digital camera and the OPTIKA PROVIEW image software. The same procedure was carried out to observe sexual products and to sex selected dissected colonies. Sclerites (selecting the largest to smallest of all possible morphological types) from different parts of the colony were mounted on stubs, coated with gold-palladium under a Leica ACE600 and observed with a Zeis EVO SEM at the General Research Services of Microscopy at the University of Seville.

The material examined for this study has been deposited in the Museu de Zoologia in Barcelona (MZB) and in the collection of the Biodiversidad y Ecología Acuática (BECA) research group at the University of Seville.

This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank Life Science Identifiers (LSID) can be resolved, and the associated information can be viewed through any standard web browser by appending the LSID to the prefix “http://zoobank.org/”. The LSID for this publication is https://zoobank.org/References/D8EEA0C1-C8F5-4BE2-9454-EFA268199A39

Class OCTOCORALLIA Haeckel, 1866

Order Scleralcyonacea McFadden, van Ofwegen and Quattrini, 2022

Superfamily Pennatuloidea Ehrenberg, 1834

Remarks

McFadden et al. (2022) recently proposed that the order Pennatulacea Verrill, 1865, which includes at least 15 families (see also López-González et al. 2022), should be abandoned in favour of a superfamily Pennatuloidea McFadden, van Ofwegen and Quattrini, 2022 in a new octocoral order named Scleralcyonacea McFadden, van Ofwegen and Quattrini, 2022. This proposal builds on previous analyses on mtMutS+ND2 (McFadden et al. 2006), further discussions on calcaxonian systematics (Williams 2019), and analyses of ultraconserved exons and loci and mtMuts (McFadden et al. 2022). The consideration of the previous order Pennatulacea as the superfamily Pennatuloidea is retained here to avoid the paraphyly of the recently described order Scleralcyonacea (see McFadden et al. 2022). However, from a nomenclatural point of view, once the taxonomic category order is proposed to be removed for this grouping, the subordinate category superfamily (a rank of the family group) is regulated under the ICZN (1999) according to the Article 36.1 Statement of the Principle of Coordination applied to family-group names. “A name established for a taxon at any rank in the family group is deemed to have been simultaneously established for nominal taxa at all other ranks in the family group; all these taxa have the same type genus, and their names are formed from the stem of the name of the type genus [Art. 29.3] with appropriate change of suffix [Art. 34.1]. The name has the same authorship and date at every rank.” This means that the authorship of this superfamily corresponds to Ehrenberg, 1834, because it was he who created the family Pennatulidae based on the type genus Pennatula according to Article 36.1, following the Principle of Coordination.

Family Funiculinidae Gray, 1860

Diagnosis (amended from Williams, 1990: 64, 69)

Colonies clavate to whip-like; with authozooids arranged bilaterally along rachis, sometimes in somewhat oblique series, rarely distally reunited, with or without spiculiferous calyces; siphonozooids on the rachis, with or without calyces; axis, present, rarely absent; sclerites, when present, three-flanged, tuberculate rods and plates. According to the current molecular knowledge, with a distinctive combination of nucleotide sequence in mtMutS (position 148-C, insertion 177-185, 203-C, 208-T, 225-T, 255-C, 358-C, 626-G, 688-A), ND2 (432-C, 537-C), and COI (273-A, 388-G) within Clade III, where the family Gyrophyllidae is also included.

Type genus

Funiculina Lamarck, 1816.

Genus Scoti a belemnon gen. nov.

Diagnosis

Funiculinid with a terminal cluster of a few (3-4) upwardly directed autozooids. Siphonozooids in two sets, in short longitudinal lines at the rachis-peduncle limit, and as a distal dome over autozooids bases. Axis absent. Sclerites mainly as large monoaxial rods with points and longitudinal furrows, scarcely tuberculate platelets, knobbly three-flanged rods, plus smooth to nearly smooth indistinctly tree-flanged needles (siphonozooid field), and ovals to oval-rods with irregular knobbly ornamentation (rachis and peduncle). According to the current molecular knowledge, with a distinctive combination of nucleotide sequence in mtMutS (position 72-T, 93-A, 117-A, 289-G, 438-G, 463-G), ND2 (209-C, 309-G), and COI (255-T, 387-T, 396-T) within Funiculinidae.

Type species

Kophobelemnon molanderi Pasternak, 1975 (=Kophobelemnon pauciflorum Molander, 1929, name unavailable by homonymy with K. pauciflorum Hickson, 1916).

Nomenclatural statement

An LSID number was obtained for the new genus: https://zoobank.org/NomenclaturalActs/27f06b9c-6c7f-4e81-bb87-4187b4184c70

Etymology

The new genus is named by combining Scotia in reference to the Scotia Sea and surrounding archipelagos, the currently known range of distribution of the type species after the works of Molander (1929), Pasternak (1975), Kushida et al. (2022) and this paper, and the Greek word belemnon (dart or javelin), a common suffix used in naming sea pen genera after the general shape of the colony. Gender neuter.

Nominal species

Scotiabelemnon pauciflorum (Molander. 1929).

Remarks on nominal species in Scotiabelemnon gen. nov.

Specimens sequenced by Kushida et al. (2022) as Pennatulacea sp. are here considered as Scotiabelemnon sp. mainly because of molecular differences in the COI mitochondrial marker and some spicular differences (Kushida and Reimer, pers. comm.) (see also phylogenetic approach part in this paper). Paradoxically, Pasternak (1975) correctly executed a nomenclatural act denoting the homonymy between Kophobelemnon pauciflorum Molander 1929 and Kophobelemnon pauciflorum Hickson, 1916, providing a new name to Molander species whose type material retained the consideration of the name bearer. However, the erection of the genus Scotiabelemnon gen. nov. and the transfer of K. molanderi (=K. pauciflorum Molander) to it make the use of K. molanderi Pasternak unnecessary, as there are no other species using the specific epithet pauciflorum in Scotiabelemnon gen. nov. Therefore, the correct name for the material examined here and the type species of this new genus should be Scotiabelemnon pauciflorum (Molander, 1929). The present paper suggests that the diversity of the genus Scotiabelemnon gen. nov. could include more than a single species. Molander’s and our present material was collected in shallower waters (200-402 m depth) than the abyssal-hadal and upper hadal (5650-6150 m depth) specimens described by Pasternak (1975) collected from South Sandwich Islands and the northern Scotia Sea, for which no molecular data are available. Moreover, even Pasternak (1975: 102-103) himself indicated that the deep-sea water colonies are significantly different from the shallow ones (они значительно отличаются от мелководных), and that tentacular sclerites are only present in the tentacular axis, unlike in Molander’s and our specimens described here, which also have sclerites in the pinnulae (see below). This is a diagnostic characteristic at the species level that is frequently used in sea pens. For these reasons, we prefer to keep the potential synonymy of Pasternak’s specimens with a question mark.

Scotiabelemnon pauciflorum (Molander, 1929) comb. nov.

Kophobelemnon pauciflorum Molander, 1929: 80.

Kophobelemnon molanderi Pasternak, 1975: 102 (substitution name for K. pauciflorum Molander, 1929 invalidated by homonymy).

(?) Kophobelemnon molanderi Pasternak, 1975: 102 (South Sandwich Islands and northern Scotia Sea specimens).

non Kophobelemnon pauciflorum Hickson, 1916: 72.

Material examined

Newly collected material: MZB 2024-3405, ANT XIX/5 - LAMPOS, Stn. 231-1, 22 Apr 2002, Agassiz trawl, 43º27.42’W 60º59.19’S, 399-402 m depth, 1 whole colony, 32 mm in length. BECA(OPEN-667), ANT XIX/5 - LAMPOS, Stn. 231-1, 22 Apr 2002, Agassiz trawl, 43º27.42’W 60º59.19’S, 399-402 m depth, 1 colony in two pieces,10 mm in length, male. BECA(OPEN-122, G-330), ANT XXIX/3 - ECOWED, Stn 240-3, 9 Mar 2013, Agassiz trawl, 62°7.05’S 60°34.12’ W 275-277 m depth, 1 whole colony, 36 mm in length. BECA(OPEN-121, G-337), ANT XXIX/3 - ECOWED, Stn 240-3, 9 Mar 2013, Agassiz trawl, 62°7.05’S 60°34.12’ W 275-277 m depth, 1 whole colony in two halves, 38 mm in length, female.

Description

Colonies varied from 10 to 38 mm in length (Figs 2A, B; 4B, D), with a terminal group of up to 4 autozooids (largest studied colony, Fig. 3A left), usually 3 autozooids from 10 mm in colony length (Figs 4C-D). Colonies without a clearly distinguishable symmetry pattern (Fig. 2B). Internal axis completely absent (Figs 3D, 4D). Peduncle (calculated in the largest colonies) 33%-40% of total colony length, not excessively bulbous in living and preserved state. Rachis composed of a stalk 60%-77% of total colony length, where relatively large autozooids appear distally (although gastrovascular cavities run internally along most of the rachis length, Fig. 3D). Lower part of the stalk cylindrical to conical widening distally. Terminal cluster of polyps distinctly swollen (Figs A-C, 3D, 4). Colony of 38 mm in total length developing numerous oocytes, up to 900 µm in diameter (Figs 3D, 5A). Colony of 10 mm in total length developing numerous spermatic cysts, up to 360 µm in diameter (Figs 4B, D, 5B, C).

Autozooids relatively short and robust (Figs 2A-C, 3B, 4), partially extended in preserved state (also in just collected material, Figs 2A-C), up to 4.5 mm in height (without tentacle length, probably much larger in fully extended state) and 4.0 mm in width, almost cylindrical, with an apparently smooth surface. Introvert folded in preserved state (Fig. 3C). Pharynx elongated and wrinkled. Tentacles relatively large and robust (Fig. 3B, C), up to 8 mm in preserved state. Pinnulae thick and relatively short, closely placed in a single line (Fig. 3B, C), up to 2.5 mm in length in preserved state, and up to 16 in number on each side of main tentacle axis.

Siphonozooids (0.2-0.5 mm in diameter) in two sets, at the basal part of rachis along intermittent longitudinal rows of 4-14 zooids (Fig. 2A, B, D), and numerous as a distal dome among the autozooid bases (Figs 2C, 4A, C).

Sclerites in all parts of the colony. Five kinds of sclerites: distinctly large monoaxial rods with numerous points and longitudinal furrows (Fig. 6A₁ , 7A₁ ), scarcely tuberculate platelets (Fig. 8A₂ ), smooth to nearly smooth indistinctly tree-flanged needles (Fig. 8A₁ ), knobbly three-flanged rods (Figs 7B₂ , C₂ ), and ovals to oval-rods with irregular knobbly ornamentation (Figs 8B, 9).

Sclerites of main tentacular axis (Fig. 6A) in two types: large monoaxial rods up to 0.42 mm in length (Fig. 6A₁ ) and elongate knobbly three-flanged rods up to 0.19 mm in length (Fig. 6A₂ ). Pinnulae with scarcely tuberculate platelets (Fig. 6B₁ ), knobbly three-flanged rods (Fig. 6B₂ ) and intermediate forms (Fig. 6B₃ ), all three up to 0.21 mm in length.

Distal part of body of autozooid with large monoaxial rods up to 0.35 mm in length (Fig. 7A₁ ) and knobbly three-flanged rod, some of them nearly smooth, up to 0.18 mm in length (Fig. 7A₂ ). Lower part (introvert) with the same sclerome but slightly smaller, large monoaxial rods up to 0.29 mm in length (Fig. 7B₁ ) and knobbly three-flanged rods up to 0.18 mm in length (Fig. 7B₂ ). Pharyngeal tissue with the same composition as polyp body wall, but slightly thinner, large monoaxial rods up to 0.27 mm in length (Fig. 7C₁ ) and knobbly three-flanged rods (somewhat transitional in appearance to large monoaxial rods) up to 0.21 mm in length (Fig. 7C₂ ).

Siphonozooids field among autozooids with smooth to nearly smooth indistinctly tree-flanged needles up to 0.4 mm in length (Fig. 8A₁ ), scarcely tuberculate platelets up to 0.24 mm in length (Fig. 8A₂ ) and short knobbly three-flanged rods up to 0.14 mm in length (Fig. 8A₃ ).

Rachis ovals to oval-rods with irregular knobbly ornamentation up to 0.18 mm in length (Fig. 8B). Peduncle sclerites similar to those from rachis but with smoother ornamentation, up to 0.16 mm in length (Fig. 9).

Colour

Living colonies dirty white to fleshy in colour (Fig. 2A, B); yellowish to orange colour in some colonies because of oocyte development. Siphonozooids as minute white spots (in both sets, those along the intermittent longitudinal rows and those at the distal dome); distal part of anthocodia (without tentacles) and pinnulae also distinctly whitish, while introvert is translucent probably because of a thinner body wall, allowing the observation by transparency of sexual products (Fig. 2B). Preserved specimens are milky to dirty white or yellowish in colour (Figs 3, 4)

Distribution

Scotiabelemnon pauciflorum (Molander, 1929) comb. nov. is distributed along the tip of Antarctic Peninsula: Bransfiel Strait (Molander 1929), South Shetland island (this paper), and near South Orkney Islands (this paper), between 200 and 402 m depth.

Phylogenetic approach

In the hypothesis using only mtMutS (Fig. 10 left), Clade III showed a strongly supported family Gyrophyllidae (bootstrap [Bst] 99%, posterior probability [PP] 1) as the sister group of a moderately-supported polytomy grouping of taxa including Kophobelemnon s.l. and Funiculina sequences (Bst 77%, PP 0.79). Regardless of whether the BI or ML method was used, four groups were clearly differentiated: three of them correspond to species identified as belonging to the genus Kophobelemnon (Bst 96 to 100%, PP 0.99 to 1, see below), while the fourth is composed of Funiculina species (Bst 74%, PP 0.88 to 1). In this tree, Funiculina sequences are reunited with a set of Kophobelemnon sequences, here informally named Kophobelemnon I (Bst 91%, PP 0.99). Another set of Kophobelemnon sequences, here informally named Kophobelemnon II, was also distinctly separated (Bst 96%, PP 0.97 to 1), and the fourth grouping reunited sequences (Bst 99%, PP 1) of those colonies identified as Kophobelemnon pauciflorum (present study) and Pennatulacea sp. sequences (GenBank accession numbers ON603970 and ON603971), all these considered here within the genus Scotiabelemnon gen. nov.

In the mtMutS+ND2+COI hypothesis (Fig. 10 right), based on all of the mitochondrial markers examined here, Clade III again showed a strongly supported family Gyrophyllidae (Bst 100%, PP 1) as the sister group of the above grouping of taxa (here considered in the reformulated family Funiculinidae), including Kophobelemnon s.l. Scotiabelemnon gen. nov. and Funiculina sequences (Bst 77%, PP 1). Regardless of whether the BI or ML method was used, four groups were clearly differentiated in Funiculinidae: two of them correspond to species identified as belonging to the genus Kophobelemnon (lineages I and II), the third reunited Scotiabelemnon gen. nov. sequences, and the fourth is composed of Funiculina sequences.

In the third hypothesis including also the nuclear 28S markers (Fig. 11), the ML and BI methods showed initially different topologies. Concerning basal clades, the ML tree (Fig. 11 right) had a similar topology to that shown in the mtMutS tree (Fig. 2 left), and the genus Gyrophyllum arose from the more basal node of Clade III. However, the BI tree (Fig. 11 left) seemed to be more affected by the poor 28S coverage in Clade III, and the genus Funiculina arose from the more basal node of Clade III. 28S sequences are available only in 6 of the 22 species in Clade III, representing 7 of the 29 sequences in our analyses. Three 28S sequences are only available from the group informally called here Kophobelemnon II, two for one of the species of Scotiabelemnon, and none from Kophobelemnon I or Funiculina. Further efforts appear to be needed to complete the coverage of 28S and to discuss a more reliable four-marker hypothesis exploring possible internal relationships in Clade III. In any case the usefulness of 28S for resolving taxa at genus and family levels is debatable because of the observed mito-nuclear incongruences (see Discussion).

Mean K2P genetic distances between groups were relatively homogeneous, Scotiabelemnon gen. nov. was 1.7% distant from Funiculina but 2.1% distant from all three other groupings (Kophobelemnon I, Kophobelemnon II and Gyrophyllum). The K2P mean genetic distances within groups were 0.0% (Gyrophyllum), 0.4% (Funiculina), 0.2% (Kophobelemnon I), 0.9% (Kophobelemnon II) and 0.0% (Scotiabelemnon gen. nov.).

The MtMutS sequence of Scotiabelemnon pauciflorum comb. nov. was 0.0% distant from the homologous sequence from colonies identified as Pennatulacea sp. (GenBank Accession numbers for mtMutS ON603970 and ON603971), denoting that, at least, they all belong to the same genus. Differences in the COI sequences suggest that more than one species of Scotiabelemon gen. nov. could be present in the study area.

The family Kophobelemnidae was described by Gray (1860) as the tribe Kophobelemnonieae to accommodate club-shaped colonies with polyps in longitudinal rows on the ventral surface, at the time to include only the genus Kophobelemnon.

From a morphological point of view, the family Kophobelemnidae included three genera: Kophobelemnon Asbjørnsen, 1856, Sclerobelemnon Kölliker, 1872 and Malacobelemnon Tixier-Durivault, 1966 (Williams 1995, McFadden et al. 2024). Because of the relative simplicity and poorly defined morphological characters (cylindrical to clavate colonies, with or without sclerites, autozooids in longitudinal rows with bilateral symmetry but not always clearly defined) the first molecular analyses that included more than the type genus (Kophobelemnon) suggested that this family is more of a taxonomic repository than a natural unit.

McFadden et al. (2006) first included sequences of Kophobelemnon and Sclerobelemnon species in a single tree (McFadden et al. 2006: Fig. 2), showing a large genetic distance between the two genera, the former aligned with Gyrophyllum (later named as Clade III) and the latter related to Pteroeides (later named as Clade I). Subsequent phylogenetic studies focusing on pennatulaceans added more sequences but showed the same scenario, a phylogenetic tree with four main clades in which these two genera are widely separated (Dolan et al. 2013, Kushida and Reimer 2019, García-Cárdenas et al. 2020). Kushida and Reimer (2019) showed that Veretillid genera and Sclerobelemnon formed a clade, with robust support from analyses utilizing the mtMutS and ND2 region. Furthermore, López-González and Drewery (2022), in a four molecular marker phylogenetic analysis, stated that “Sclerobelemnon merges among veretilid genera, thus the family Veretilidae can only be considered monophyletic if Sclerobelemnon is included within it”, a placement that is also supported by phylogenetic hypothesis based on mtMutS and ND2 analysed separately (López-González and Drewery 2022: 218 and Fig. 15-16).

The genus Malacobelemnon Tixier-Durivault, 1966 is still a question mark. This genus was based on M. stephensoni Tixier-Durivalt, 1966, a species with few usable morphological characteristics that has never been collected again, and no molecular information of it is available. More recently, López-González et al. (2009) described a second species in this genus because of its morphological simplicity, and trying to be conservative in the systematics of sea pens, the species was included in this genus as M. daytoni López-González, Gili and Fuentes, 2009. However, the first amplifications of newly collected material of M. daytoni showed it to be highly divergent (López-González unpublished data), while a deeper whole mitochondrial study is in preparation (Figueroa and López-González, in prep.). Thus, even considering that M. daytoni was correctly assigned to the genus Malacobelemnon, the pertinence of this genus to the old family Kophobelemnidae or to the here reformulated Funiculinidae remains doubtful and tentative.

López-González et al. (2022) concluded Clade III to be composed of three families: Funiculinidae (with the genus Funiculina), Kophobelemnidae (including at that time only the genus Kophobelemnon) and Gyrophyllidae (with the genus Gyrophyllum). All these genera were the type genus of their respective families, although some sequences in the genus Kophobelemnon were distinctly divergent. According to the current molecular knowledge and available sequenced genera and species (Dolan et al. 2013, Kushida and Reimer 2019, García-Cárdenas et al. 2019, López-González and Drewery 2022, López-González et al. 2022, Hogan et al. 2023), it seems that evolutionary hypotheses can be divided into at least four, this paper representing a fifth because of the introduction of Scotiabelemnon gen. nov. (Fig. 12):

Hypothesis 1 had Gyrophyllum as the basal taxon of Clade III and nested Kophobelemnon and Funiculina (Dolan et al. 2013, Kushida and Reimer 2019, García-Cárdenas et al. 2020 [ML]). When Kophobelemnon I and II can be differentiated (mainly thanks to the amount of sequenced specimens included in the analysis), Funiculina is usually grouped with Kophobelemnon I distally. This topology is observed when mtMutS is analysed alone (by BI or ML) and when mtMutS is part of a concatenated dataset of mitochondrial markers only (by BI).

Hypothesis 2 did not support the large Clade III (García-Cárdenas et al. 2019 [BI]). 28S is included in a concatenated four-marker data set analysed by BI.

Hypothesis 3 had Funiculina as the basal taxon of Clade III and nested Kophobelemnon and Gyrophyllum (López-González and Drewery 2022, López-González et al. 2022). Gyrophyllum is grouped with Kophobellemnon II sequences distally. This occurs when 28S is included as the fourth marker.

Hypothesis 4 had Gyrophyllum as the basal taxon of Clade III and nested Funiculina and Kophobelemnon sequences (López-González and Drewery 2022). This topology occurs when only the three mitochondrial markers are analysed. Kophobelemon I and II are distinct and located distally with respect to Funiculina. A derived version of this hypothesis (as 4* in Fig. 12) is in Kushida et al. (2022), where a couple of divergent sequences based on specimens from the South Sandwich Islands (Antarctic waters) as Pennatulacea sp. were included in the analysis (P in Fig. 12). Pennatulacea sp. are related to Kophobelemnon II distally. These sequences are considered in this paper to belong to Scotiabelemnon gen. nov.

Hogan et al. (2023) utilized only Kophobelemnon and Funiculina for phylogenetic trees utilizing mitochondrial genome, and it was difficult to discuss the phylogenetic hypothesis within their Clade III as Gyrophyllum and potential Scotiabelelmnon gen. nov. data were not available.

All phylogenetic trees in this study supported Hypothesis 1 (derived as 1* by the inclusion of Scotiabelemnon gen. nov. sequences). Gyrophyllum is the basal taxon of Clade III and nested Kophobelemnon s.l., Scotiabelemnon gen. nov. and Funiculina (this paper, see Figs 10, 12). Basal relationships of these last three genera are still uncertain, but Funiculina is always grouped with Kophobelemnon I distally, while it is poorly suggested that Scotiabelemnon gen. nov. could be their sister group.

The genera Funiculina and Gyrophyllum have an unstable location when 28S is included in the analyses, likely affected by evolutionary differences between the mito-nuclear region (see Quattrini et al. 2023, Hogan et al. 2023) and the poor coverage of this marker in our datamatrix (only 6 of the 26 in Clade III are available). The inclusion of nuclear 28S in the mitochondrial marker data set in combination with the inference model used resulted in different taxa combinations (BI placed as basal Gyrophyllum when 28S is not included, but Funiculina when 28S is incorporated in the data matrix, while ML usually places Gyrophyllum as the basal taxon in Clade III, with and without 28S in the data matrix). Incomplete taxonomic and molecular coverage, especially at the 28S marker, probably also affects the results for a reliable four-marker hypothesis (see above).

In this paper, which includes the broadest coverage of Kophobelemnon and Funiculina sequences, the pennate colony forms of the family Gyrophylliadae are clearly differentiated from the flagelliform and clavate colony forms (Funiculina and Kophobelemnon s.l.+Scotiabelemnon gen nov., respectively). The sister group of Gyrophyllidae is a polytomy (mtMutS) bringing together four subclades, three of them with relatively low basal resolution, which includes sequences from Kophobelemnon, Scotiabelemnon gen. nov. and Funiculina. Among these four groups, the sequences initially attributed to Kophobelemnon represent three of them, the fourth being Funiculina sequences. One of these Kophobelemnon clusters is described here as the genus Scotiabelemnon gen. nov., while the other two are simply named here as Kophobelemnon I and II.

The observed stronger support of Kophobelemnon I clade to the Funiculina clade (Bst 91, PP 0.99 in mtMutS; Bst 92, PP 0.93 in mtMutS+ND2+COI;) than to Kophobelemnon II and Scotiabelemnon gen. nov. raises additional discussions concerning the delimitation of the families involved. Surprisingly, Scotiabelemnon gen. nov. is closer to Funiculina spp. (in K2P genetic distance) than to other sequenced specimens attributed to the genus Kophobelemnon I or II. In any event, Gyrophyllidae and Funiculinidae become monophyletic, but Kophobelemnidae (Kophobelemnon s.l. + Scotiabelemnon gen. nov.) become paraphyletic, Kophobelemnon I being closer to Funiculina spp. that to other putative species of Kophobelemnon in Kophobelemnon II. The proposal of this paper to avoid paraphyly of Kophobelemnidae is the reunion of Funiculina, Kophobelemnon s.l. and Scotiabelemnon gen. nov. in a single family, Funiculinidae. It is therefore proposed to consider only two families in Clade III, Gyrophyllidae and Funicunilidae.

The family name, Funiculinidae Gray 1860 must be used, applying the principle of priority (ICZN 1999: Article 52.3). In 1860 Gray inserted into our sea pen literature as tribe names Funiculininae (Gray 1860: 20) and Kophobelemninae (Gray 1860: 23), and these names were elevated to family rank by Gray in 1870, as Funiculinidae (Gray 1870: 12) and Kophobelemnidae (Gray 1870: 27), using the appropriate suffix to accommodate the original names to the new taxonomic category.

The current conception of this family is the result of the merge of the monotypic family Funiculinidae with the family Kophobelemnidae (see Williams 1990 for previous separate morphological diagnoses) because of the close molecular relationships of their type genera Funiculina, and Kophobelemnon s.l., as well as the new genus proposed here, Scotiabelemnon gen. nov. As molecular systematics comes to redefine the boundaries between taxonomic categories previously considered stable, morphological diagnoses of several octocoral families become less morphologically diagnostic than ever, sometimes being merely informative about the morphological diversity of the taxa included in them (see McFadden et al. 2022).

In the present paper we propose a diagnosis for the family Funiculinidae that complements morphological information with those molecular characteristics that currently distinguish the two families considered within Clade III, Funiculinidae Gray 1860 and Gyrophyllidae López-González, Drewery and Williams 2022. Obviously, as with any type of characteristics included in the diagnosis of a taxon, morphological and molecular information must be updated to cover the variability of the taxa considered in it at any given time.

In the present study, we considered with certainty three nominal genera in this reformulated family Funiculinidae: Funiculina, Kophobelemnon and Scotiabelenon gen. nov. Considering that the genus Kophobelemnon is in urgent need of revision, and that at least an additional genus name (new or resurrected from old literature) will be added once a reliable revision of these species is carried out.

Other genera previously considered in Kophobelemnidae, such as the genera Sclerobelemnon and Malacobelemnon, have been discussed above. Sequences of Sclerobelemnon are strongly attracted to genera in the family Veretillidae Herklots, 1858 (see García-Cárdenas and López-González 2020, López-González and Drewery 2022), so its morphological features are not considered in this reformulated family Funiculinidae. As discussed above, the inclusion of Malacobelemnon in the family Funiculinidae is tentative until reliable molecular information on its type species becomes available.

The genus Kophobelemnon was erected by Asbjørnsen (1856) for the species Kophobelemnon muelleri Asbjørnsen, 1856, a junior synonym of Pennatula stellifera Müller, 1776. Once diversity of pennatuloid genera was expanded during the 19th century, P. stellifera was subsequently transferred from Pennatula Linnaeus, 1758 to Veretillum Cuvier, 1798, Umbellularia Lamarck, 1801 and Funiculina, and was finally accommodated in Kophobelemnon (Kükenthal and Broch 1911: 224, Kükenthal 1915: 29).

From a morphological point of view, the genus Kophobelemnon is characterized by the following characters: 1) colonies elongate and cylindrical to slightly clavate or short, stout and distinctly clavate; 2) distal end rounded and knob-like or distinctly pointed; 3) rachis with bilateral symmetry throughout, although not always distinctly so; 4) axis thin, round to slightly quadrangular in cross section; 5) polyp leaves absent, autozooids up to approx. 50 in number, arranged biserially along rachis; 6) anthocodiae mostly not retractile, calyces absent; 7) siphonozooids numerous, often with minute spiculated calyces, on areas of rachis not occupied by autozooids; and 8) sclerites densely set, spindles and rods, mostly three-flanged, sometimes ornamented with tubercles (Williams 1995: 108-109).

Obviously, there is a strong morphological similarity between the colonies of the two Kophobelemnon lineages (I and II), but only one can retain the genus name. In this case we prefer not to decide here which genus name must be used for each of these two Kophobelemnon groupings yet, because these nomenclatural acts should only be carried out after a deep morphological and molecular study in which morphological features are explored and segregated and molecular differences have been established by the analysis of solitary genes or concatenated gene datasets (see Fig 12; Dolan et al. 2013, Kushida and Reimer 2019; Kushida et al. 2022; López-González and Drewery 2022; this paper). In this case, the genus name Kophoblemenon must be retained when Kophobelemnon muelleri Asbjørnsen, 1856, the type species of the genus [today recognized as Kophobellemon stelliferum (Müller, 1776)], is correctly recognized. Rather than being an easy task, this step is a challenge. Müller (1776) described in his Zoologiæ danicæ prodromus, seu animalium daniæ et norvegiæ indigenarum the species Pennatula stellifera based on material from an imprecise locality from the northern seas, whereas in both main Kophobelemnon lineages (I and II) specimens from the northeastern Atlantic have been sequenced. At this time, the sequenced specimen Kophobelemnon sp. 1-ED seems to be the geographically closest to the type locality of K. muelleri (=Kophobellemon stelliferum), although this fact may not be a decisive criterion. This sequence (MutS GenBank accession number KF313837) is placed in the here unofficially named Kophobelemnon I lineage. This lineage is also present in other localities of the northeastern Atlantic and western Pacific, while the Kophobelemnon II lineage is present in the northeastern Atlantic and the western and eastern Pacific.

Morphological features of Scotiablemnon molanderi comb. nov. can be used for a diagnostic description of a new genus, even though these features were previously considered to be included within the morphological variability of the genus Kophobelemnon. Apart from the molecular differences, the genus Scotiabelemnon gen. nov. is differentiated from Kophobelemnon s.l. in the following set of characters: 1) siphonozooids in two sets, short longitudinal lines over rachis-peduncle limit and as a distal dome among the few autozooid bases; 2) autozooids placed distally and upwardly directed (including gastrovascular cavities), and 3) absence of axis. As chronologically described above, the genus Kophobelemnon was quickly recognized as a para- or polyphyletic taxon (Dolan et al. 2013, Kushida and Reimer 2019, García-Cárdenas et al. 2020, López-González and Drewery 2022, López-González et al. 2022). Moreover, the last phylogenetic hypotheses by Kushida et al. (2022) included sequences of additional species attributed here to the genus Scotiabelemnon gen. nov. This fact suggests that the diversity of this last genus is still unknown, and its placement in the phylogenetic trees reinforces the current paraphyletic nature of the genus Kophobelemnon.

In the current scenario, Linnaean taxonomy (understood as the discipline that regulates how to name a hierarchical set of subordinate taxonomic categories) and phylogenetic hypotheses (generated after the analysis of an ever-increasing amount of molecular information) should find a conciliatory space to achieve a better and useful view of the tree of life. Undoubtedly, the addition of new sequences, even when taxa are only identified at the genus level, greatly helps to outline the limits of higher taxonomic units and to recognize para- or polyphyletic taxa (species, genus, or families) previously named on the basis of morphological characters. The nomenclatural problems produced by the inconsistencies observed in the phylogenetic trees should be progressively resolved through the complete description of the affected taxa in a combined morphological and molecular study, which is ultimately called integrative taxonomy (Di Camillo et al. 2018, Gómez Daglio and Dawson 2019).