New genera and species from the Equatorial Pacific provide phylogenetic insights into deep-sea Polynoidae (Annelida)

Polynoidae contains ~900 species within 18 subfamilies, some of them restricted to the deep sea. Macellicephalinae is the most diverse among these deep-sea subfamilies. In the abyssal Equatorial Pacific Ocean, the biodiversity of benthic communities is at stake in the Clarion-Clipperton Fracture Zone (CCFZ) owing to increased industrial interest in polymetallic nodules. The records of polychaetes in this region are scarce. Data gathered during the JPI Oceans cruise SO239 made a significant contribution to fill this gap, with five different localities sampled between 4000 and 5000 m depth. Benthic samples collected using an epibenthic sledge or a remotely operated vehicle resulted in a large collection of polynoids. The aims of this study are as follows: (1) to describe new species of deep-sea polynoids using morphology and molecular data (COI, 16S and 18S); and (2) to evaluate the monophyly of Macellicephalinae. Based on molecular and morphological phylogenetic analyses, ten subfamilies are synonymized with Macellicephalinae in order to create a homogeneous clade determined by the absence of lateral antennae. Within this clade, the Anantennata clade was well supported, being determined by the absence of a median antenna. Furthermore, 17 new species and four new genera are described, highlighting the high diversity hidden in the deep. A taxonomic key for the 37 valid genera of the subfamily Macellicephalinae is provided.


INTRODUCTION
The family Polynoidae Kinberg, 1856 is one of six families called scale worms (Aphroditiformia). With ~900 species belonging to 167 genera in 18 subfamilies (Wehe, 2006;Norlinder et al., 2012;Read & Fauchald, 2018), polynoids are the most diverse polychaetes in number of genera and the second most diverse in number of species (Hutchings, 2000;Wehe, 2006;Read & Fauchald, 2018). They are errant worms with wide-ranging distribution, from shallow intertidal Branchipolynoinae has four species in one genus and Branchiplicatinae one species in one genus. The high number of monotypic subfamilies and genera observed in deep-sea polynoids raises concerns and suggests either that subfamilies do not represent monophyletic groups and should be reviewed in a phylogenetic context or that there is a high diversity of as yet undiscovered species within these subfamilies/genera, the discovery of which might support the monophyly of these groupings. Based only on morphological data, the paraphyly of Polynoidae subfamilies has already been suggested by Rouse & Pleijel (2001), who stressed the need to investigate these in a phylogenetic framework.
The phylogeny and evolutionary origin of the subfamily Macellicephalinae within polynoids is still unclear. Phylogenetic studies that included Polynoidae to date were mainly concerned with phylogeny of Annelida (Rouse & Fauchald, 1997;Zrzavy et al., 2009) or the deeper relationships of scale worms (Aphroditiformia) at the family level (Wiklund et al., 2005;Norlinder et al., 2012;Gonzalez et al., 2018). With regard to Macellicephalinae, Uschakov (1977Uschakov ( , 1982 suggested the presence of derived character states in the worms from this subfamily (e.g. short body and reduction of jaws), hypothesizing that they originated from a common ancestor with genus Bathymoorea Pettibone, 1967. Two studies using combined morphological and molecular data have already suggested the paraphyly of the subfamily Macellicephalinae. Norlinder et al. (2012) showed a well-supported clade (three representatives of deep-sea subfamilies) with members of the subfamilies Macellicephalinae, Branchinotogluminae and Branchipolynoinae. However, owing to the limited number of sequences from deep-sea subfamilies available, no hypothesis about their relationships was developed. In a larger, but still limited, number of seven representative species of deep-sea subfamilies, Gonzalez et al. (2017Gonzalez et al. ( , 2018 recovered two main clades: (1) Macellicephalinae, Branchinotogluminae and Branchipolynoinae, similar to Norlinder et al. (2012); and (2) a clade composed only of members of Macellicephalinae. The lack of knowledge about deep-sea polynoids inevitably reflects the small number of samples, particularly DNA friendly, from the deep sea. The relative larger body size, low density and high mobility of polynoids reduce the probability of adequate sampling with the widely used quantitative methods (i.e. box cores, mega-cores; e.g. De Smet et al., 2017). However, a greater abundance of worms can be sampled with qualitative methods, such as an epibenthic sledge (EBS; e.g. Brandt & Schnack, 1999;Janssen et al., 2015), which can capture mobile epibenthos, such as polynoids.
The Clarion-Clipperton Fracture Zone (CCFZ) is the largest polymetallic nodule field in the world, with ~6 million km 2 of seabed lying between 4000 and 6000 m depth. Mining these nodules might directly impact 300-800 km 2 yr −1 of the seafloor, and sediment plume re-deposition might indirectly increase the footprint of mining by a factor two to five (Glover & Smith, 2003). The polychaete fauna in the CCFZ is highly diverse (up to 113 taxa per 0.25 m 2 ) but remains largely undescribed (5-10% of worms identified to named species; Glover et al., 2002). The polynoids of CCFZ are virtually unknown.
A large collection of deep-sea polynoids collected from epibenthic sledges and a remotely operated vehicle deployed during the JPI Oceans cruise SO239 led to the discovery of ~80 molecular operational taxonomic units (MOTUs) and, for the first time ever, enabled us to provide insights into the phylogeny of this poorly understood group. In this context, the aims of this work were as follows: (1) to describe 17 new species of deep-sea polynoids using morphology together with molecular data in most cases (COI,16S and 18S); and (2) to assess the monophyly of the subfamily Macellicephalinae.

Clarion-Clipperton FraCture Zone
The CCFZ is a vast area, ~6 million km 2 , in the Equatorial Pacific Ocean, bounded by the Clarion Fracture to the north, the Clipperton Fracture to the south, the Exclusive Economic Zone (EEZ) of Kiribati to the west and Mexico to the east. This area is composed of abyssal hills and seamounts of major commercial interest because of the presence of polymetallic nodules. The CCFZ could hold 34 billion metric tons of manganese (Morgan, 2000). The International Seabed Authority (ISA), established by the United Nations Convention on the Law of the Sea (UNCLOS) is in charge of managing deepsea mineral resources beyond national jurisdictions and protecting the marine environment, including fauna and flora, against any pollution or another hazard (Articles 145, 156, UNCLOS;Lodge et al., 2014). Up to now, the ISA has concluded 16 contracts for the exploration of polymetallic nodules in the CCFZ. In view of future exploitation of nodules, contractors are required to carry out surveys and inventories of mega-, macro-and meiofauna associated with nodules fields and to evaluate the potential impact(s) of mining on the environment and biota in their contract area (Lodge et al., 2014). In addition, at the regional scale, the ISA approved in 2012 an Environmental Management Plan that includes the designation of nine 'Areas of Particular Environmental Interest' (APEIs).
In this context, the European project 'Managing Impacts of Deep-seA reSource exploitation' (MIDAS) and the pilot action 'Ecological aspects of deep-sea mining' of the Joint Programming Initiative Healthy and Productive Seas and Oceans (JPI Oceans) are aimed at providing sound scientific bases to assess and manage the impact of potential mining activities on deep-sea ecosystems.
Within the framework of JPI Oceans pilot action, the SO239 cruise took place from 9 March to 30 April 2015, on board the RV Sonne (Chief Scientist: Professor Dr Pedro Martinez). The exploration areas of four contractors and an APEI were sampled between 4000 and 5000 m depth (Fig. 1). The exploration areas were licensed by the ISA to the Federal Institute for Geosciences and Natural Resources of Germany (BGR) sponsored by Germany; the InterOceanMetal Joint Organization (IOM) sponsored by Bulgaria, Cuba, Czech Republic, Poland, Russian Federation and Slovakia; the G-TEC Sea Mineral Resources NV (GSR) sponsored by Belgium; and the Institut français de recherche pour l'exploitation de la mer (Ifremer) sponsored by France. The APEI number 3 (APEI#3) is administrated by ISA as part of the regional environmental plan for the CCFZ (Fig. 1). Within each area, macrofaunal samples were collected using an epibenthic sledge (Brenke, 2005) which consists of a supraand epibenthic net with cod ends of 300 µm each and an opening and closing mechanism; or the remotely operated vehicle (ROV) Kiel 6000, which has different sampling tools. One of them, the biobox, is a large box in which specimens collected with the manipulator arm were stored.

Morphology
The specimens were live sorted, photographed with a Canon EOS 700D, fixed in cold (−20 °C) 80% ethanol and kept at −20 °C. In the laboratory, a small piece of tissue (e.g. a few parapodia, an end of the body or tissue from the pharynx when everted) was sampled and fixed in cold 96% ethanol for molecular studies (see DNA extraction, amplification, sequencing and alignment). Preserved specimens were examined under a Leica M125 stereomicroscope and a Nikon Eclipse E400 microscope and photographed with a Nikon DS-Ri 2 camera. Body length and width (at segment 4, without parapodia) were measured using Leica LAS Interactive Measurements. Drawings were carried out on stacked (overlapped and aligned) pictures on an iPad Pro using an Apple pencil and Graphic App, with formatting using Adobe Illustrator and Photoshop to assemble them into plates. Specimens examined with scanning electron microscopy (SEM) were prepared by dehydration with three immersions of 15 min each in 96% ethanol, critical-point drying, and being covered with gold and photographed using the Quanta200 FEI (Ifremer). Morphological characters were coded either from original descriptions or redescriptions; from Gonzalez et al., (2018), who examined specimens of Gesiella jameensis, Halosydnella australis, Harmothoe rarispina and Pelagomacellicephala iliffei; or from morphological examination of type specimens described in this study. From Gonzalez et al. (2018), 50 out of 87 coded characters were used in the present study. In addition, 24 characters were newly included in order to represent the morphological characters of deep-sea subfamilies. Seventy-four characters (Supporting Information, Appendix S1) were included in a morphological data matrix (Supporting Information, Appendix S2). However, the characters that were regularly missing in the literature or were doubtful (e.g. number or size of pharyngeal papillae) were not included in the matrix.
Type specimens were deposited at the Muséum National d'Histoire Naturelle de Paris (MNHN, France) and at the Natural History Museum, London (NHM, UK). Additional material was mostly archived at the MNHN, and a few specimens remain in the private collection of P.B., as recorded in the 'Material examined' sections and Table 1.
Dna extraCtion, aMpliFiCation, sequenCing anD alignMent Subsamples for DNA analysis were removed from live specimens, placed in 96% ethanol and frozen at −20 °C. Extraction of DNA was done with a NucleoSpin Tissue (Macherey-Nagel) kit, following the protocol supplied by the manufacturers. Approximately 450 bp of 16S, 700 bp of COI (cytochrome c oxidase subunit I) and 1800 bp of 18S genes were amplified using the following primers: Ann16SF and 16SbrH for 16S (Palumbi, 1996;Sjölin, et al., 2005); polyLCO, polyHCO, LCO1490 and HCO2198 for COI (Folmer et al., 1994;Carr et al., 2011); and 18SA, 18SB, 620F and 1324R for 18S (Cohen et al., 1998;Medlin et al., 1998;Nygren & Sundberg, 2003) for 18S. The polymerase chain reaction (PCR), with 25 µL mixtures containing: 5 µL of Green GoTaq Flexi Buffer (final concentration of 1×), 2.5 µL of MgCl 2 solution (final concentration of 2.5 mM), 0.5 µL of PCR nucleotide mix (final concentration of 0.2 mM of each dNTP), 9.875 µL of nuclease-free water, 2.5 μl of each primer (final concentration of 1 µM), 2 µL template DNA and 0.125 U of GoTaq G2 Flexi DNA Polymerase (Promega). The temperature profile was as follows: 95 °C for 240 s, followed by [   by 4 °C. The PCR products that produced light bands after electrophoresis on 1% agarose gel were sent to the MacroGen Europe Laboratory in Amsterdam (The Netherlands) to obtain sequences, using the same set of primers as used for the PCR.
Overlapping sequence (forward and reverse) fragments were merged into consensus sequences using Geneious Pro 8. 1.7 2005. For COI, the sequences were translated into amino acid alignments and checked for stop codons to avoid pseudogenes. The minimal length coverage was 344 bp for 16S, 531 bp (only one with 345 bp) for COI and 1215 bp (two with < 800 bp) for 18S.
The sequences were blasted in GenBank in order to check for the presence of contamination. Each set of genes was aligned separately using the plugins MAAFT (Katoh et al., 2002) for 16S and 18S, and MUSCLE (Edgar, 2004) for COI. All sequences obtained in this study have been deposited in GenBank (Table 1; http:// www.ncbi.nlm.nih.gov/genbank/).

phylogenetiC analyses
Bayesian and maximum likelihood phylogenetic analyses were run with three different datasets (Table 2): a molecular dataset (MDS; 65 taxa) based only on DNA sequences, including COI, 16S and 18S genes; a combined dataset (CDS; 65 taxa) based on the MDS dataset with the addition of morphological data for the species having a complete description for the type species; and a morphological dataset (128 taxa) based only on the 74 coded characters. A dataset with all molecular and morphological data was not possible because of the lack of sequence data for most of the species.
The model used for all genes was GTR+G (Gonzalez et al., 2018). Each gene set was combined in a partitioned dataset with SequenceMatrix (Vaidya et al., 2011). The Bayesian phylogenetic analysis was performed with MrBayes v.3.2.6 on XSEDE (Ronquist et al., 2012), following standard methods with 60 000 000 generations. At every 1000 generation one chain was sampled; in the end, 25% were discarded as burn-in. At the end, the convergence chain runs were assessed using TRACER v. 1.7.1 (Rambaut et al., 2018). The support of nodes is given as Bayesian posterior probability (BPP) values. The maximum likelihood was performed using Randomized Axelerated Maximum Likelihood (RAxML v.8.2.10;Stamatakis, 2014) on XSEDE with rapid bootstrapping (1000 iterations). The support of nodes is given as maximum likelihood bootstrap (MLB) values. MrBayes and RAxML were computed in CIPRES Science Gateway (Miller et al., 2010). Morphological data were analysed using the MK model (Lewis, 2001) in both analyses. The most parsimonious analysis was performed using PAUP v.4.0a (build 161) based only on the morphological data matrix (Supporting Information, Appendix S2), with all observations weighted equally and multistate characters treated as unordered, and the following string was run a few times: hsearch enforce = no start = stepwise addseq = random nreps = 20 000 nchuck = 5 chuckscore = 1 nbest = all. Characters were plotted on the strict consensus tree based on the most parsimonious trees using MacClade v.4 (Maddison & Maddison, 2005). The tree files were analysed with FigTree v.1.4.2 (available at http:// tree.bio.ed.ac.uk/software/figtree/).        Barnich & Fiege (2003); Gonzalez et al. (2018) genetiC DivergenCe Molecular operational taxonomic units were recognized using a threshold of 97% similarity between COI sequences (Hebert et al., 2003a, b). The similarity of sequences within species is given in the 'Genetic data' sections in order to compare with this general threshold. Furthermore, the average evolutionary divergence over sequence pairs was calculated within species (intraspecific variation) and between species (interspecific variation) using the Kimura two-parameter (K2P;Kimura, 1980) model in MEGA7 (Kumar et al., 2016). This analysis allowed for the comparison of genetic distance, mainly between closely related  The morphological analysis yielded five most parsimonious trees with 396 state changes (strict consensus tree in Fig. 3) rearranged 9 139 993 959 times, with a consistency index of 0.24 and retention index of 0.75. The morphological analysis showed the same patterns as the MDS and CDS analyses. Most polynoids with lateral antennae were subdivided into two main groups without a unique synapomorphy assembling them (grey lines; Fig. 3). Among these groups, only the subfamily Admetellinae (clade c1; i n a e a n d Vampiropolynoinae being determined by the unique synapomorphy absence of lateral antennae (character 11: 0). Within this group, five subgroups were supported, each with a unique synapomorphy (Fig. 3). Clade d1 included Bruunilla species, being determined by the presence of a lower lip with a winglike structure (character 6: 1). Clade d2 included only members of the subfamily Macellicephaloidinae, being determined by the presence of dissimilar dorsal and ventral jaws (character 40: 1). Clade d3 included the Anantennata clade with all members of the subfamilies Polaruschakovinae, Bathyedithinae and the genus Hodor gen. nov. (not assigned to a current subfamily), being determined by the absence of a median antenna (character 7: 0). Clade d4 included Bathyeliasona species, being determined by the presence of very wide neurochaetae (character 60: 1). Clade d5 included only members of the subfamily Branchipolynoinae (Fig. 3), being determined by the absence of ceratophores of the median antenna (character 9: 0).         , 4, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 26, 29, 32 and 33. Parapodia subbiramous, notopodia shorter than neuropodia; noto-and neuropodia with elongate acicular lobe; tips of noto-and neuroaciculae not penetrating epidermis. Notochaetae numerous, with spinous rows; neurochaetae numerous.
Remarks: Pettibone (1997) has erected this subfamily for Eulagisca McIntosh, 1885 and Pareulagisca Pettibone, 1997. According to Pettibone, the prostomium having lateral antennae inserted terminally or subterminally is a character shared with the subfamilies Lepidonotinae and Lepidastheniinae, but the parapodia shape is distinctive in Eulagiscinae. Indeed, in Lepidonotinae and Lepidastheniinae the notopodia are vestigial to moderately developed, whereas in Eulagiscinae they are well developed (Wehe, 2006). We suggest that Bathymoorea should be included in Eulagiscinae because it has lateral antennae inserted subterminally and well-developed notopodia. Thus, the description has been emended in order to include the characters of Bathymoorea, such as: absence of chaetae on tentacular segment (present in Eulagisca and Pareulagisca as opposed to present or absent in Bathymoorea), the absence of a nuchal fold (present in Eulagisca and Pareulagisca as opposed to absent in Bathymoorea), number and shape of eyes (two small pairs in Eulagisca and Pareulagisca as opposed to one pair of large eyes in Bathymoorea) and shape of dorsal tubercles (bulbous/nodular in Eulagisca and Pareulagisca as opposed to lamelliform/ inflated in Bathymoorea).
Remarks: Pettibone (1967) erected the genera Bathyadmetella Pettibone, 1967 and Bathymoorea and emended Admetella McIntosh, 1885 based on bathyal specimens from the Central and North-eastern Pacific. According to Pettibone (1967) and Uschakov (1982), these genera share some similarities: prostomium with paired large eyes, lepidonotoid-like ceratophores of the lateral antennae, bulbous facial tubercles, long neuroacicular lobes and neurochaetae of one type, numerous, long and flattened. Uschakov (1977) created the subfamily Admetellinae to include Admetella and Bathyadmetella, both possessing antennal scales or sheaths. As pointed out by Uschakov (1982), the position of Bathymoorea, without scales or sheaths on the antennae, remained unknown until now. The genus is here emended in order to include the presence of chaetae on the tentacular segment and nephridial papillae from segment 5 or 6 observed in Bathymoorea lucasi sp. nov.

Description (based on holotype and paratypes):
Holotype complete, 9.95 mm long and 1.35 mm wide for 27 segments (including tentacular segment), dorsoventrally flattened, not tapering posteriorly; live specimen coloration pale yellow body and prostomium; ethanol-preserved specimen with pale yellow body (Fig. 4A), dark dots present on middle and posterior margin of prostomium, dark spots sparsely covering mid-ventrum of body and dorsal surface of notopodia, styles of lateral antennae and ventral cirri brownish medially to distally, nephridial papillae brownish. Prostomium bilobed, wider than long, with large pair of whitish ocular areas (Fig. 4H, A). Median and lateral antennae present; ceratophore of median antenna large, bulbous, inserted near anterior margin, style missing; lateral antennae inserted on anterior extension of prostomium subterminally to ceratophore of median antenna, styles smooth, tapering, short (about one-quarter length of palps). Palps smooth, tapering distally to thin tips, short (reaching to segment 4; Fig. 4H). Facial tubercle present, bulbous.
Morphological variation: Specimens with 25, 26, 27 and 29 segments were found. The form of the prostomial appendages, shape of parapodia and form of chaetae were similar to those of the types. However, specimens with 25 and 29 segments possessed 12 and 13 pairs of elytrophores, respectively. Differences are probably linked to size or growth/development stages of the animals, because DNA confirmed that all specimens belonged to the same species. Furthermore, variation in the first occurrence of nephridial papillae is also linked to size; in animals with 25 segments the nephridial papilla started from segment 5 or 6, whereas in animals with > 25 segments the nephridial papilla always started from segment 5. Worms also differed in having short and long dorsal cirri in the anterior part of the body, but in the posterior body the observed dorsal cirri were always long. Loss and regeneration could explain the very short dorsal cirri observed in the anterior body.
Remarks: Bathymoorea has contained a single species, Bathymoorea renotubulata (Moore, 1910) since its erection by Pettibone (1967). The specimens found in the CCFZ are similar to Bathymoorea renotubulata as follows: short body, large ocular areas, similar prostomial shape and form of distally flattened neurochaetae with faint spinous rows, which appear bidentate in lateral view as opposed to unidentate in Bathymoorea lucasi sp. nov. Bathymoorea lucasi sp. nov. differs from Bathymoorea renotubulata in having fewer segments (≤ 29), very short lateral antennae, short palps, presence of acicula and notochaetae on the tentacular segment, short neuropodial lobes and very reduced (bulbous) nephridial papillae. In comparison, Bathymoorea renotubulata has 33 segments, long antennae, long palps, achaetous tentaculophores, elongate neuropodial lobes and elongate nephridial papillae. With regard to chaetae, notochaetae are robust with well developed rows of spines in Bathymoorea lucasi sp. nov. instead of the delicate and fine spines in Bathymoorea renotubulata.
Etymology: This species is dedicated to Lucas Lisboa, cousin of P.B., for his friendship.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S, respectively sharing at least 98.6, 99.3 and 100% of genetic material between the specimens. The average K2P distance for intraspecific variation was 0.8% for COI and 0.2% for 16S.
Distribution: Based on the material examined (nine specimens), this species has a restricted distribution within the Clarion-Clipperton Fracture Zone, being sampled at two stations in APEI#3 nodules area (type locality).
Ecological notes: During the sampling at station 189, 11 Hexactinellida sponges were sampled together with a few ophiuroids and alcyonaceans, all conditioned in the biobox of the ROV. At station 200, six Hexactinellida sponges were sampled together with anthipatharians, crinoids, hydrozoans, ophiuroids and tunicates. The specimens of the new species Bathymoorea lucasi sp. nov. were found by sieving water from the biobox, which indicates a possible commensalism with the sponges.

Morphological variation:
The specimens differing in number of segments (16 and 18) otherwise showed great morphological similarities, including: size of appendages, form of prostomium and its appendages, first occurrence of nephridial papillae, form of parapodia and type of chaetae. However, animals with 16 segments had eight pairs of elytrophores instead of nine pairs, probably related to size.

Remarks:
No other genera of Macellicephalinae show neuropodia with lanceolate pre-chaetal lobes and falcate neurochaetae differing between segment 2 and subsequent segments. In Uncopolynoinae the species Uncopolynoe corallicola likewise shows strongly curved hooks present in anterior segments (Wehe, 2006). This character might be linked to the life history of these worms, which are always found living on alcyonarian corals. Furthermore, Parahololepidella greeffi (Augener, 1918) is commensal on antipatharians (Britayev et al., 2014) and shows neurochaetae rather similar to Abyssarya acus gen. nov., sp. nov. This evidence reinforces a possible commensal relationship between Abyssarya acus gen. nov., sp. nov. and corals recovered from the same sampling biobox.
Etymology: The species name came from Latin 'ăcŭs' meaning 'needle'. It refers to modified neurochaetae present on segment 2 similar to a 'crochet needle'.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S, respectively sharing at least 99.1, 99.8 and 100% of genetic material between the specimens. The average K2P distance for intraspecific variation was 0.5% for COI and 0.1% for 16S.
Distribution: Only five specimens were sampled and all at a single station within the Clarion-Clipperton Fracture Zone in APEI#3 seamount area (type locality).
Ecological notes: These worms were found in the water sieved from the ROV biobox at station 212, which contained sponges (Hexactinellidae), alcyonaceans, antipatharians and pennatulacean corals, in addition to ophiuroids. Abyssarya acus gen. nov., sp. nov. is likely to be commensal with one of these taxa, and more studies in the area are needed to identify the host.

Remarks:
The diagnosis of genus Bathyedithia is emended to include a character observed in the species described below and the species Bathyedithia tuberculata Levenstein, 1981: ten pairs of elytrophores; as well as with characters observed in the new species: the presence of a terminal anus and the number of pairs of pharyngeal papillae. Description (based on holotype): Holotype complete, 3.81 mm long and 0.71 mm wide for 20 segments (including tentacular segment), dorsoventrally flattened, posteriorly tapering; live specimen bluish, transparent (Fig. 6A); ethanol-preserved specimen pale white.
Tentacular segment fused to prostomium, with a pair of short lobes, inserted laterally and slightly ventral to prostomium; without acicula or chaetae; tentaculophores distinct, bulbous, equal sized; styles smooth, tapering into thin tips, short (reaching to segment 4), dorsal and ventral tentacular cirri of similar length (Fig. 6B). Pharynx dissected, with seven pairs of subtriangular distal papillae of similar size; two pairs of jaws with main fang, serrated margin (one pair of jaws with nine to 11 teeth and the other pair with 13 teeth; Fig. 6C). Second segment with elytrophores, subbiramous parapodia, chaetae and ventral cirri.
Segment 6-8 without any structures. Ventral cirri smooth, tapering into thin tips, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (longer than tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia, style short (shorter than tip of neuroacicular lobe).

Remarks:
The new species is very close to Polaruschakov species having seven pairs of pharyngeal papillae, but even closer to Bathyedithia in having serrated jaws and rounded palpophores. Bathyedithia retierei sp. nov. differs from the two other species of Bathyedithia in having seven pairs of pharyngeal papillae and lanceolate neuropodia. Furthermore, in Bathyedithia retierei sp. nov. the nephridial papillae are present on segments 10 and 11, whereas in Bathyedithia tuberculata they are present from segment 7 to the end of the body, and they are absent in Bathyedithia berkeleyi.
Etymology: This species is dedicated to Professor Christian Retière (Muséum National d'Histoire Naturelle, Dinard, France) for his many contributions to French benthic research.
Genetic data: DNA sequencing for this species was successful for 16S and 18S (only 720 bp) but not for COI.
Remarks: The diagnosis of the genus is emended to add the following characters observed in the new species described below: nuchal organs present and notochaetae stouter than neurochaetae.  Description (based on holotype and paratype): Holotype complete, 9.86 mm long and 1.88 mm wide for 17 segments (including tentacular segment), dorsoventrally slightly flattened, not tapering posteriorly; live specimen iridescent, purplish in colour dorso-anteriorly, prostomium whitish (Fig. 7A, B); ethanol-preserved specimen pale yellow, prostomium whitish; golden aciculae.
Prostomium bilobed, about as wide as long, lobes pronounced, anteriorly extending into slender, long frontal filaments; median notch between prostomial lobes moderately wide and deep; eyes absent; a pair of comma-shaped, pigmented nuchal organs in a slight depression present on mid anterior prostomial lobes (Fig. 7B, C). Median antenna present, lateral antennae absent; ceratophore of median antenna cylindrical, long (not surpassing anterior end of frontal filaments), inserted in median notch, style missing. Palps smooth, tapering, very long (reaching to segment 8; Fig. 7C). Facial tubercle absent.
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style short (much shorter than tip of neuroacicular lobe); in subsequent segments inserted subdistally on neuropodia (Fig. 7F), styles short (approaching tip of neuroacicular lobe); in last segment, style longer than in preceding segments. Parapodia subbiramous; notopodia shorter than neuropodia (Fig. 7F). Notopodia arising from the dorsum, as two thickened ridges; notopodia subtriangular, tapering into long acicular lobe, tip of notoacicula not penetrating epidermis. Neuropodia large, rectangular to subtriangular, tapering into long acicular lobe, tip of neuroacicula not penetrating epidermis. The last pair of parapodia similar to preceding ones. Notochaetae of two types: (1) very few (four observed), short, stout, slightly curved with distinct spinous rows on convex side, with blunt tips (Fig. 7G); and (2) moderate in number (12 observed), long to very long, slender, slightly curved with distinct, well-developed, clear spinous rows, occasionally one of those notochaetae can be stouter and straight (Fig. 7H), with blunt tips; notochaetae stouter and narrower than neurochaetae. Neurochaetae of two types: (1) upper group, variable in number (three to 19 observed), long, distally very wide, flattened, serrated along both margins, with pointed tips (Fig. 7I); and (2) middle and lower group, numerous (12-35 observed), long to short, distally wide, flattened, serrated along both margins, with pointed tips (Fig. 7J), with a kind of central rib distally present on some specimens (Fig. 7J).

Morphological variation:
The specimens found in the CCFZ are morphologically very similar, having 15 and 17 segments. The common morphological characters included: shape of prostomium and form of prostomial appendages, shape of parapodia and types of chaetae. However, the specimen with 15 segments showed some differences probably related to growth: seven pairs of elytrophores, median notch less prominent, longer ventral cirri, longer palps, pygidium not clearly bilobed and not enclosed by the last parapodia.
R e m a rk s : O n l y t h r e e s p e c i e s b e l o n g i n g t o Bathyeliasona are known: Bathyeliasona kirkegaardi (Uschakov, 1971), Bathyeliasona abyssicola (Fauvel, Hartman, 1967. These species were reviewed by Pettibone (1976), who separated them using characters such as the number of segments, type of notochaetae and development of notopodia/neuropodia in the last segments. Although not mentioned by Pettibone (1976), differences in the shape of pygidium can be highlighted from her drawings, which can also be used to separate species. The species Bathyeliasona abyssicola and Bathyeliasona nigra have 18 segments (including tentacular segment), whereas Bathyeliasona kirkegaardi has 17 segments, like Bathyeliasona mariaae sp. nov. Additional similarities between the latter two species are: shape of jaws, shape of pygidium and notopodial lobes shorter than neuropodia on the last segments. However, while Bathyeliasona mariaae sp. nov. has prostomial lobes anteriorly extending into long frontal filaments that reach the distal end of the ceratophore of the median antenna, in Bathyeliasona kirkegaardi the prostomial lobes are anteriorly rounded and the frontal filaments are minute and filiform, shorter than the ceratophore of the median antenna. Furthermore, in Bathyeliasona kirkegaardi, Bathyeliasona abyssicola and Bathyeliasona nigra the length of the ventral cirri is much shorter than the neuroacicular tip and the notochaetae are of one type only (stout with distinct spinous rows), whereas in Bathyeliasona mariaae sp. nov. the length of the ventral cirri is approaching the neuroacicula tip and the notochaetae are of two types (stouter with distinct spinous rows and slender with developed spinous rows). Pettibone (1976) described all the Bathyeliasona species with notochaetae more slender than neurochaetae, but in Bathyeliasona mariaae sp. nov. one notochaeta per notopodium can be stouter (and always narrower) than the neurochaetae. Furthermore, in Bathyeliasona mariaae sp. nov. the middle and lower neurochaetae show a kind of central rib distally that gives the neurochaetae a bilimbate appearance.

1913) and Bathyeliasona nigra
Etymology: This species is dedicated to Maria Silva, mother of P.B., for her love.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S, with all three genes sharing 100% of genetic material between the specimens. The average K2P distance for intraspecific variation was 0.0% for both COI and 16S.
Remarks: The diagnosis of genus Bathyfauvelia is emended to include characters observed in the new species described below. These include: absence of frontal filaments and presence of serrated jaws. Pettibone (1976) considered valid only two species in this genus, but a few inconsistencies were observed from the drawings of the Arctic and Mediterranean specimens of Bathyfauvelia affinis (Pettibone, 1976: 36-37, figs 21, 22). The specimen from the Arctic Basin shows an elongated acicular lobe in the tentacular segment without chaetae and shorter prostomium lobes, whereas the specimen from the Mediterranean Sea shows a shorter acicular lobe in the tentacular segment with chaetae and slightly longer prostomium lobes.
Morphological variation: Only one specimen shows an adult size, confirmed by the presence of nephridial papillae. All the other specimens appear to be juveniles sharing many similarities with the adult: long palps (reaching to segment 4-5) and form of notochaetae and neurochaetae. However, the prostomium in juveniles shows a wider notch, and peaks are poorly developed.
Remarks: Currently, only two species belonging to Bathyfauvelia are valid: Bathyfauvelia affinis and Bathyfauvelia grandelytris (Levenstein, 1975). Both species share the presence of pointed cephalic peaks (anterior end), presence of cirriform dorsal tubercle and only one type of neurochaetae, whereas the new species described here presents a similar prostomium shape and dorsal tubercle but has two types of neurochaeta present (Table 3). Bathyfauvelia glacigena sp. nov.
is very similar to Bathyfauvelia ignigena sp. nov., and the two species can be easily confused (Table 3), but evidence from DNA shows that they are distinct species (Fig. 2). This was confirmed by the average K2P distance between them (14.0% for COI and 7.8% for 16S). Adult specimens of Bathyfauvelia glacigena sp. nov. have 18 segments, prostomial lobes anteriorly tapering to pointed cephalic peaks, palps slightly longer (reaching to segment 5-6) and the last ventral cirri about as long as the neuropodial lobe. In contrast, Bathyfauvelia ignigena sp. nov. has 19 segments, prostomial lobes anteriorly tapering to rounded cephalic peaks, palps slightly shorter (reaching to segment 3-4) and the last ventral cirri slightly longer than the neuropodial lobe. Only one adult specimen of Bathyfauvelia glacigena sp. nov. with 18 segments has been observed, and it is unclear whether this is the maximal number of segments for the species. More specimens are needed to confirm this character. It should be noted that prostomial lobes anteriorly tapering to blunt cephalic peaks were also observed in a few juvenile specimens of Bathyfauvelia glacigena sp. nov., suggesting that this character is ontogenetic and should be used only to differentiate adult specimens. The length of palps and ventral cirri on the last segment appear to be more consistent characters for separating these two species. Both species have overlapping distributions in the IOM and GSR license areas. In addition, Bathyfauvelia glacigena sp. nov. was also sampled in the Ifremer license area, and Bathyfauvelia ignigena sp. nov. was also sampled in APEI#3.
Etymology: The species name glacigena means 'iceborn', which is composed by borrowing from the Latin word 'glăcĭēs' meaning 'ice' and the Greek word 'gennó, γεννώ' meaning 'born'. It refers to white ganglia like ice.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S, respectively sharing at least 98.6, 99.5 and 100% of genetic material between the specimens. The average K2P distance for intraspecific variation was 1.6% for COI and 0.2% for 16S.
Distribution: Based on the material examined (six specimens), this species has a wide distribution within the Clarion-Clipperton Fracture Zone, being sampled in the IOM, GSR (type locality) and Ifremer license areas.  Description (based on holotype and paratypes): Holotype complete, 5.65 mm long and 0.80 mm wide for 19 segments (including tentacular segment), dorsoventrally flattened, slightly tapering posteriorly; colour of live animal not known; ethanol-preserved specimen pale white, prostomium transparent with two large white patches interiorly (Fig. 9A).
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (longer than tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia but basally on neuropodia of posterior segments (Fig. 9L), styles very short (shorter than tip of neuroacicular lobe); last ventral cirri longer than neuropodial lobe of same segment. Parapodia subbiramous; notopodia shorter than neuropodia (Fig. 9L). Notopodia subtriangular, tapering into long acicular lobe, tip of notoacicula not penetrating epidermis. Neuropodia large, subtriangular, tapering into long acicular lobe, tip of neuroacicula not penetrating epidermis. Notochaetae of two types: (1) few (six to eight observed), short to long, stout, slightly curved with distinct spinous rows on curved side, with blunt tips (Fig. 9D, M); and (2) moderate in number (13 observed), long to very long, slender, slightly curved with distinct, well-developed spinous rows, with blunt tips (Fig. 9E, N); notochaetae stouter than neurochaetae. Neurochaetae of two types: (1) upper group, moderate in number (~12 observed), long to very long, distally flattened to concave, serrated along both margins, with pointed tips (Fig. 9F, O); and (2) middle and lower group, moderate in number (25 observed), long to short, stouter, distally concave to folded, with spines (two to 19 observed) along both margins, with gently curved pointed tips (Fig. 9G, P-R); the lower neurochaetae in fascicle much shorter (Fig. 9Q, R), with fewer lateral spines (two or three observed), not present on segments 2-4; in last segment neurochaetae can be very thin.
Morphological variation: Most of the specimens have an adult size with 19 segments, with few morphological variations. Only one adult specimen presents a very short palp (MNHN-IA-PNT 76, reaching segment 2), which might be regenerating. The only juvenile specimen, with 12 segments, already has the two types of both notochaetae and neurochaetae present in adults but less numerous. However, the prostomial lobes (slightly wider), the prostomial peaks (poorly developed) and the dorsal tubercles (poorly developed) differ from the adults.
Remarks: Bathyfauvelia ignigena sp. nov. differs from Bathyfauvelia glacigena sp. nov. in having 19 segments, rounded cephalic peaks, slightly shorter palps (reaching to segment 3-4) and slightly longer ventral cirri than neuropodial lobes on the last parapodia. See Remarks on Bathyfauvelia glacigena sp. nov. and Table 3 for more details.
Etymology: Species named from the 'ignĭgĕna', a poetical epithet of Bacchus meaning 'fire-born', which is composed by borrowing from the Latin word 'ignis' meaning 'fire' and the Greek word 'gennó, γεννώ' meaning 'born'.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S (only one specimen). The specimens shared ≥ 99.3% for COI and 100% and for 16S. The average K2P distance for intraspecific variation was 0.4% for COI and 0.0% for 16S.
Distribution: Based on the material examined (five specimens), this species has a wide distribution within the Clarion-Clipperton Fracture Zone, being sampled in IOM (type locality), GSR, Ifremer (Janssen et al., 2015) and APEI#3 areas.
Tentacular segment fused to prostomium, with a pair of short lobes, inserted laterally and slightly below prostomium; tentaculophores small; styles smooth, tapering; dorsal tentacular style very short (reaching segment 2); ventral tentacular style long (reaching around segment 4). Pharynx dissected, with seven pairs of distal subtriangular papillae; two pairs of jaws, each one with one main fang, outer margin with a secondary very small tooth (pointed or blunt). Lips of mouth in a bulbous projection when pharynx not everted.
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style short (much shorter than tip of neuroacicular lobe); missing in subsequent segments. Parapodia subbiramous; notopodia shorter than neuropodia. Notopodia tapering into long acicular lobe, tip of notoacicula not penetrating epidermis. Neuropodia rectangular, tapering into long acicular lobe, tip of neuroacicula not penetrating epidermis.
Nephridial papillae not seen. Ventral keel present on last segments.

Remarks:
The sampled specimens belong to Bathypolaria and share many characters with the single species described in this genus (Bathypolaria carinata): short body (15 segments), flattened notochaetae and ventral keel at the posterior end. However, the specimens are in poor condition and did not allow for thorough species differentiation or formal description. Based on genetic material, two MOTUs can be found in the area. Furthermore, the average K2P distance for interspecies variation was 23.4% for COI and 15.3% for 16S, confirming the presence of two species. However, Bathypolaria sp. 173 was widely distributed (BGR, IOM and GSR), whereas Bathypolaria sp. 608 was restricted to APEI#3 area.
Genetic data: DNA sequencing for this species was successful for COI (only one specimen), 16S and 18S. The specimens shared at least 99.5% and 99.9% of genetic material in 16S and 18S genes, respectively. The average K2P distance for intraspecific variation was 0.3% for 16S.
Distribution: Based on the material examined (three specimens), this species has a wide distribution within the Clarion-Clipperton Fracture Zone, being sampled in BGR, IOM and GSR license areas. Description (based on all specimens): Fragile worms, all damaged; parapodia, prostomial and parapodial appendages often missing. Body dorsoventrally cylindrical; live specimen with body surface translucent to milky, prostomium whitish, pharynx red; ethanol-preserved specimens with body surface pale white, prostomium white, pharynx internally red, longitudinal ventral whitish line (Fig. 10B).
Tentacular segment fused to prostomium, with a pair of short lobes, inserted laterally and slightly below prostomium; tentaculophores distinct, small; dorsal tentacular style very short (reaching segment 2), smooth, tapering. Pharynx everted, with seven pairs of distal subtriangular papillae, equal sized (Fig. 10B); two pairs of jaws, each one with one main fang, outer margin with a secondary very small tooth (pointed or blunt; Fig. 10D). Lips of mouth in a bulbous projection when pharynx not everted.
Nephridial papillae not seen. Anal cirri lost, scars not seen. Ventral keel present on last segments.
Remarks: See Remarks on Bathypolaria sp. 173 for more details.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S, respectively sharing at least 99.6, 99.5 and 99.9% of genetic material between the specimens. The average K2P distance for intraspecific variation was 0.2% for COI and 0.1% for 16S.
Distribution: Based on the material examined (five specimens), this species has a restricted distribution within the Clarion-Clipperton Fracture Zone, being sampled at two stations in APEI#3 area.
Remarks: This genus was first classified as a Fauveliopsidae Hartman, 1971, but the unique specimen (without any chaetae) was reviewed by Pettibone (1979), who transferred it to Macellicephalinae. The genus Bruunilla is easily recognized by the presence of a wing-like structure located ventrally on segments 1-3. The generic diagnosis is emended here in order to include characters observed in the new species described below, such as: presence of frontal filaments, neuropodial papillae and types of notochaetae and neurochaetae. Pettibone (1979) did not mention the presence of neuropodial papillae, although Hartman (1971) drew it as a short papilla on segment 7 (Hartman, 1971(Hartman, : 1412 fig. 3). This kind of papilla has been observed before in species of Diplaconotum Loshamn, 1981(Loshamn, 1981 fig. 5C), as a small, short papilla also on the neuropodia. However, this latter genus belongs to Polaruschakovinae, not having median or lateral antennae. This fact could explain the closer relationship between Bruunilla and taxa without antennae. The genus Bruunilla has been monotypic for almost four decades since its erection (Pettibone, 1979). Increased exploration of the deep ocean has led to the discovery of several new forms that possess wing-like structures on their ventrum, a hallmark of this genus. One species is described here, but several others (seven MOTUs) were recognized by DNA only. Unfortunately, those specimens were in too poor condition to allow for a morphological description (data not presented). Our findings suggest that there is a much greater diversity of this genus in the deep sea than previously thought. Description (based on holotype): Holotype complete, 3.00 mm long and 0.69 mm wide for 17 segments (including tentacular segment), dorsoventrally flattened, slightly tapering posteriorly; live specimen pale white, slightly translucent; ethanol-preserved specimen pale yellow.
Prostomium bilobed, wider than long, lobes not pronounced, anteriorly rounded; with short, ovoid frontal filaments; median notch between prostomial lobes narrow and shallow (Fig. 11A); eyes absent. Median antenna present, lateral antennae absent; ceratophore of median antenna bulbous, very short (shorter than anterior end of frontal filaments), inserted near posterior margin of prostomium; style smooth, tapering, long (reaching segment 7). Palps smooth, tapering, short (reaching segment 4; Fig. 11A). Facial tubercle not seen. A pair of wing-like structures on ventral side, like a prolongation of lower lip, separated longitudinally until segment 3; median notch between lobes narrow and deep; each lobe oval, with blunt tips, with most longitudinal folds anteriorly, and most horizontal folds posteriorly, partially covering first four or five segments (Fig. 11B, G).
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style smooth, tapering, long (much longer than tip of neuroacicular lobe); in subsequent segments inserted subdistally on neuropodia (Fig. 11C), styles long (about as long as tip of neuroacicular lobe).
Remarks: Bruunilla natalensis is the only species described in this genus until now. Bruunilla nealae sp. nov. is unique in having frontal filaments and long neuropodial papillae on segments 11-14, which are absent in Bruunilla natalensis (papilla maybe present on segment 7; see Remarks on Bruunilla). In addition, the blunt tips of the wing-like structure in Bruunilla nealae sp. nov. differ from the protruding tips of the wing-like structure in Bruunilla natalensis. Furthermore, the average K2P distance between Bruunilla nealae sp. nov. and Bruunilla sp. 692 was very high (27.3% for 16S). Etymology: This species is dedicated to Lenka Neal (Natural History Museum, London) for her friendship, discussion and comments during this project.
Genetic data: Gene sequencing for this species was successful for 16S and 18S (only 1200 bp) but not for COI.
Distribution: Only one specimen was sampled at a single station within the Clarion-Clipperton Fracture Zone in Ifremer license area (type locality). Description: Complete, 3.16 mm long and 0.55 mm wide, dorsoventrally flattened; colour of live animal not known; ethanol-preserved specimen pale yellow; poor condition, most chaetae and parapodia missing (Fig. 11H).
Prostomium bilobed, wider than long, lobes not pronounced, anteriorly rounded; with ovoid frontal filaments; median notch between prostomial lobes narrow and shallow (Fig. 11H); eyes absent. Median antenna present, lateral antennae absent; ceratophore of median antenna bulbous, very short (shorter than anterior end of frontal filaments), inserted near posterior margin of prostomium, style smooth, tapering, long. Palps missing. A pair of wing-like structures on ventral side (left one broken), like a prolongation of lower lip, separated longitudinally until segment 3; each lobe oval, blunt, with mostly longitudinal folds anteriorly and posteriorly, partly covering at least four segments.
Tentacular segment with a pair of short lobes, inserted laterally and slightly below prostomium; tentaculophores small, equal sized; styles missing. Pharynx dissected with pharyngeal papillae not possible to count; two pairs of jaws, each with main fang and serrated margins (18-22 teeth); gradually bigger distally. Second segment with elytrophores, subbiramous parapodia, with chaetae and ventral cirri.
Ventral cirri present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style smooth, tapering, long (much longer than tip of neuroacicular lobe); missing in subsequent segments.
Notochaetae missing. Neurochaetae long to short, with distal parts flattened to concave, serrated along both margins, with pointed to blunt tips.

Remarks:
The specimen was too damaged to be described formally as a new species but was successfully sequenced for COI, 16S and 18S.
Genetic data: DNA sequencing for this specimen was successful for COI, 16S and 18S.
Distribution: Only one specimen was sampled at a single station within the Clarion-Clipperton Fracture Zone in APEI#3 area. hodor gen. nov.

Remarks:
The presence of large palpophores and serrated jaws should place this genus as a member of Bathyedithinae, but the presence of seven pairs of pharyngeal papillae is a character present in Polaruschakovinae. The genus can thus not be assigned to any of the current subfamilies, which supports synonymizing most of the deepsea subfamilies (see Discussion). The distinctive characters of Hodor gen. nov., which separate the genus from Bathyedithia, are: seven pairs of similarsized distal papillae and the presence of stouter modified neurochaetae on segments 3-7, whereas Bathyedithia has nine pairs of distal papillae, with the middle one elongated, and lacks modified neurochaetae. In this study, the combined molecular and morphological phylogenetic analysis placed Hodor gen. nov. closer to Polaruschakov, which is consistent with their morphology, given that in both genera the median and lateral antennae are absent. Description (based on holotype): Holotype complete, 11.35 mm long and 1.68 mm wide for 24 segments (including tentacular segment), dorsoventrally flattened, posteriorly tapering; colour of live animal not known; ethanol-preserved specimen pale white (Fig. 12A).
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (longer than tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia (Fig. 12D), style short (shorter than tip of neuroacicular lobe).
Parapodia subbiramous; notopodia reduced, much short than neuropodia (Fig. 12D). Notopodia arising from the dorsum as two thickened ridges; narrow, subtriangular, tapering into long acicular lobe, tip of notoacicula not penetrating epidermis. Neuropodia large, rectangular to subtriangular, tapering into long acicular lobe, tip of neuroacicula not penetrating epidermis. Notochaetae variable in number (one to 16 observed), long, slender, slightly curved with distinct, faint spinous rows on convex side, with blunt tips preceded by subdistally smooth margin (Fig. 12E); notochaetae more slender than neurochaetae. Neurochaetae of two types: (1) moderate in number (12-28 observed), long to very long, distally flattened to concave, serrated along both margins, with pointed tips (Fig. 12F); and (2) middle and lower group on segments 3-7 modified, moderate in number (16 observed), stouter, short to long, distally flattened to concave, with coarse spines along both margins, spines concentrated basally and well spaced later, middle part with smooth margins (most of or less of their length), subdistally smooth or with one or two spines, blunt tip, tip margin slightly lighter (Fig. 12G, H).
Remarks: Hodor hodor gen. nov., sp. nov. presents many similarities with Hodor anduril gen. nov., sp. nov., including the number of segments, prostomium and parapodial shape, chaetae and number of nephridial papillae pairs. However, the palps in Hodor hodor gen. nov., sp. nov. are longer (reaching segment 9) than the palps in Hodor anduril gen. nov., sp. nov. (reaching segment 2). Furthermore, the average K2P distance between these species was 14.2% for COI and 6.1% for 16S. Interestingly, the morphological variation observed in Hodor anduril gen. nov., sp. nov. specimens could also indicate some dissimilarity within Hodor hodor gen. nov., sp. nov. specimens, which will necessitate care in the identification of any further specimens. In addition, both species have overlapping distribution in APEI#3.
Etymology: This species is dedicated to Hodor, one of P.B.'s favourite characters in the novel 'A song of ice and fire' by George R. R. Martin.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S.
Distribution: Only one specimen was sampled at a single station within the Clarion-Clipperton Fracture Zone in APEI#3 area (type locality).
Tentacular segment well developed, with a pair of short lobes, inserted laterally and slightly below prostomium; without acicula or chaetae; tentaculophores large, bulbous, equal sized (Fig. 13A); tentacular styles missing. Second segment with elytrophores, subbiramous parapodia, chaetae and ventral cirri. Pharynx dissected, with seven pairs of subtriangular distal papillae, abruptly tapering, increasing in length gradually towards middle, with middle pair longest; two pairs of jaws with main fang, serrated margin (one pair of jaws with 15-16 teeth and the other with ten or 11 teeth; Fig. 13B).
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style missing; in subsequent segments inserted medially on neuropodia (Fig. 13C), style short (shorter than tip of neuroacicular lobe).
Parapodia subbiramous; notopodia reduced, much shorter than neuropodia (Fig. 13C). Notopodia arising from the dorsum as two thickened ridges; narrow, subtriangular, tapering into long acicular lobe, tip of notoacicula not penetrating epidermis. Neuropodia large, rectangular to subtriangular, tapering into long acicular lobe, tip of neuroacicula not penetrating epidermis. Notochaetae variable in number (one to 13 observed), long, slender, slightly curved with distinct, faint spinous rows on convex side, with blunt tips preceded by smooth margin (Fig. 13D); notochaetae more slender than neurochaetae. Neurochaetae of two types: (1) moderate in number (13-21 observed), long, distally flattened to concave, serrated along both margins, with pointed tips (Fig. 13E); and (2) middle and lower group on segments 3-7 modified, moderate in number (16 observed), stouter, short to long, distally flattened to concave, with coarse spines along both margins, spines concentrated basally and well spaced later, middle part with smooth margins (most of or less of their length), subdistally smooth or with one or two spines, blunt tips, tip margin slightly lighter (Fig. 13F, G).
Pygidium rounded, not enclosed by last segment; with terminal anus. Anal cirri lost, scars not seen.
Morphological variation: Holotype and paratype agree in many characters (e.g. number of segments, prostomium and parapodial shape, and form of chaetae) and their DNA (see Genetic data) but they show two important differences. In the holotype, the palps are very short (reaching segment 2), the segments 6 and 8 have swollen dorsal structure, and the nephridial papillae are present on segments 11-13, whereas in the paratype the palps are very long (reaching segment 11), the swollen dorsal structure is absent on segments 6 and 8, and the nephridial papillae are present on segments 10-13. These differences might be linked to sexual dimorphism.
R e m a rk s : D i f f e r e n c e s i n t h e f i r s t s e g m e n t with nephridial papillae have been observed in Branchipolynoe seepensis (Jollivet et al., 2000). They suggested that male specimens present one pair of nephridial papillae on segment 11, whereas female specimens present two pairs on segments 10 and 11. Thus, we believe that the differences observed between the holotype and paratype of Hodor anduril gen. nov., sp. nov. are linked to sexual dimorphism. In view of the morphological variation observed in this species, more attention needs to be paid to this genus. The long palps in the paratype resemble those of Hodor hodor gen. nov., sp. nov., but the first segment with nephridial papillae is different. In Hodor anduril gen. nov., sp. nov., nephridial papillae start on segment 10, whereas in Hodor hodor gen. nov., sp. nov. nephridial papillae start on segment 11. Furthermore, the presence of a swollen structure on anterior segments seems to indicate a reproductive feature, which is perhaps temporary.
Etymology: The species name is derived from the sword named 'andúril' meaning 'Flame of the West' and belonging to Aragorn in the novel 'The lord of the rings' by J. R. R. Tolkien. It refers to the sword-like modified neurochaetae present in this species. D, notochaeta with faint spinous rows, segment 6. E, neurochaeta, segment 6. F, stouter neurochaeta, lower group, segment 6. G, stouter neurochaeta, lower group, segment 6. Abbreviations: el, elytrophore; sw, swollen dorsal structure.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S. Holotype and paratype shared 100% of genetic material in COI, 16S and 18S. The average K2P distance for intraspecific variation was 0.0% for both COI and 16S.
Distribution: Only two specimens were sampled and both at a single station within the Clarion-Clipperton Fracture Zone in APEI#3 area (type locality).
Nephridial papillae on segments 10, 11 and 12, enlarged, rounded. Last segment very reduced; notoacicula and neuroacicula about the same size. Pygidium bulbous, not enclosed by last segment; with dorsal anus. Anal cirri lost, scars not seen.
Morphological variation: All specimens show great similarities in the shape of the prostomium, form of the chaetae and size of the neuro-and notoacicula on the last segment. However, the paratype has slightly longer frontal filaments.
Remarks: Macellicephala clarionensis sp. nov. is very close to Macellicephala parvafauces sp. nov., with both having truncated prostomial lobes and very faint spinous rows on the notochaetae. However, in Macellicephala clarionensis sp. nov. the median notch between the prostomial lobes is very narrow, the median facial tubercle is not so pronounced, and the neurochaetae present minute bidentate tips, whereas Macellicephala parvafauces sp. nov. has a moderately wide median notch, a well-developed median facial tubercle and neurochaetae with hooked tips. Macellicephala clarionensis sp. nov. is the only species within Macellicephala having minute bidentate tips ( Table 4). The average K2P distance between the two new species was high (26.9% for COI and 25.8% for 16S).
Etymology: The term 'clarionensis' refers to the area of distribution; APEI#3 is crossed by the Clarion Fracture. During the SO239 cruise, the area was sampled for the first time ever.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S. Specimens shared ≥ 99.3% of genetic material in 16S and 100% of genetic material in COI or 18S. The average K2P distance for intraspecific variation was 0.0% for COI and 0.6% for 16S.
Distribution: Only three specimens were sampled and all at a single station within the Clarion-Clipperton Fracture Zone in APEI#3 area (type locality).
Ecological notes: The specimens were found in the water sieved from the ROV biobox at station 212, which contained several species of sponges (hexactinellid), alcyonaceans, antipatharians and pennatulacean corals, in addition to ophiuroids. Furthermore, spicules were found inserted in the body of the worms, suggesting an artefact of sieving or a commensal life with sponges. More studies in the area are needed to find out which species is the host.

Description (based on holotype and paratype):
Holotype complete, 11.82 mm long and 1.50 mm wide for 18 segments, slightly dorsoventrally flattened, not tapering posteriorly; live specimen whitish anteriorly and slightly brownish posteriorly; ethanol-preserved specimen pale yellow throughout (Fig. 15A, C). Body surface smooth.
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (longer than tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia (in middle body slightly subdistally on neuropodia), style short (shorter than tip of neuroacicular lobe), style longer than neuropodia lobe in last segment. Parapodia subbiramous; notopodia shorter than neuropodia (Fig. 15F). Dorsal ridges absent. Notopodia subtriangular, tapering into long acicular lobe, tip of notoacicula not penetrating epidermis. Neuropodia large, subtriangular, tapering into long acicular lobe, tip of neuroacicula not penetrating epidermis. Notochaetae moderate in number (15-20 observed), short to long, distally straight to slightly curved, with very faint spinous rows, with pointed tips (Fig. 15G); some with slightly hooked tip; notochaetae as stout as neurochaetae. Neurochaetae of two types: (1) upper group, moderate in number (~20 observed), long to very long, distally flattened to concave, with very faint spinous rows along both margins, with pointed tips (Fig. 15H), laterally appearing with hooked tips; and (2) middle and lower groups, moderate in number (~20 observed) long to short, distally flattened to concave with very faint spinous rows on both sides but distally not touching the margins, with minute hooked tips (Fig. 15I).

Morphological variation:
The specimens found in the CCFZ are highly similar. The only difference between them concerns the chaetae. In the holotype, the chaetae are slightly wilted, giving an impression of flatness, whereas in the paratype the chaetae have the usual appearance. This could be an artefact of the fixation/ conservation process.
Remarks: Macellicephala parvafauces sp. nov. is closer to Macellicephala violacea (Levinsen, 1886) and Macellicephala australis Wu & Wang, 1987, having a smooth body surface and neurochaetae with hooked tips (Table 4). However, the styles of the dorsal cirri in Macellicephala violacea are shorter than the tip of neuropodial lobe, whereas in Macellicephala australis they are longer, as in Macellicephala parvafauces sp. nov. This new species also has prostomial lobes anteriorly truncated and conical dorsal tubercles, whereas Macellicephala australis has prostomial lobes anteriorly rounded and no dorsal tubercles.
Etymology: The species name came from the Latin 'parvă fauces' meaning 'small jaws'. It refers to the relatively smaller jaws of this species.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S. The holotype and paratype shared 98.6% of genetic material in COI and 100% in 16S or 18S. The average K2P distance for intraspecific variation was 1.4% for COI and 0.0% for 16S.
Distribution: Based on the material examined (two specimens), this species has a wide distribution within the Clarion-Clipperton Fracture Zone, being sampled in APEI#3 (type locality) and GSR license areas.
Remarks: Currently, eight species have been described (Pettibone, 1989b) in this relatively speciose genus, presenting such particular morphological characters as: modification in prostomium shape, dorsal jaws fused, absence of notochaetae and very short to very long notoacicula. The genus description is emended here in order to include the characters observed in the new species below: very short and slender notoacicula.  Description (based on holotype and paratypes): Holotype complete, 2.69 mm long and 0.66 mm wide for 15 segments (including tentacular segment), slightly dorsoventrally flattened, slightly tapering posteriorly; live specimen bluish; ethanol-preserved pale brown (Fig. 16A), with brownish bands subbasally on palps and medially on dorsal cirri; prostomium whitish.
Prostomium bilobed but continuously fused, wider than long, lobes not pronounced anteriorly but extending to ventral side (Fig. 16A-D); frontal filaments small, inserted anteriorly ( Fig. 16A-C); median notch between prostomial lobes absent; eyes absent. Median antenna present, lateral antennae absent; ceratophore of median antenna cylindrical, short (shorter than distance between ceratophore insertion and posterior margin of prostomium), inserted anteriormedially on prostomium; style missing in holotype; in paratype (MNHN-IA-TYPE 1833) style smooth, tapering into thin tips, long (reaching segment 5). Palps smooth, tapering, long (reaching segment 4-5), inserted ventrally, directed ventroposteriorly (Fig. 16C, D). Pharynx not everted in holotype; dissected in paratype (MNHN-IA-TYPE 1834), with pharyngeal papillae not possible to count; two pairs of jaws, dorsal jaw fused, each dorsal jaw with two main blunt teeth, innermost one with a smaller, pointed secondary tooth, margin smooth (Fig. 16E); each ventral jaw with one main blunt tooth, outer margin serrated (four to five small teeth), with two smaller secondary teeth, pointed on each side, margin smooth and short (Fig. 16F). Bilobed facial tubercle, very small, located between palps and upper lip (Fig. 16D). In most specimens, bilobed facial tubercle difficult to see but made visible using forceps to move palps apart.
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style missing in holotype; in paratypes (MNHN-IA-TYPE 1833 and MNHN-IA-TYPE 1835), style short (shorter than tip of neuroacicular lobe); in subsequent segments inserted distally on neuropodia, styles short on segments 3 and 4 (shorter than tip of neuroacicular lobe) and long from segment 5 (longer than tip of neuroacicular lobe).

Morphological variation:
The specimens vary in the number of segments from 11 (MNHN-IA-PNT 81) to 16 (MNHN-IA-PNT 79) but share many of the morphological characters described above except for the number of pairs of elytra and the presence of nephridial papillae. The specimen with 11 segments presents five pairs of elytrophores and no nephridial papillae; these characters are probably linked to growth.
Methyl Green staining pattern: The paratype (NHMUK 2018.25351) stains slightly but uniformly all along the body (not shown). Two densely stained spots are observed anteriorly on prostomium, located close to the median ceratophore.  Levenstein, 1983, shows notopodia shorter than neuropodia, with thin notoacicula (Pettibone, 1989b). However, the new species Macellicephaloides moustachu sp. nov. differs from Macellicephaloides improvisa in having up to 16 segments and elongate cirrophores with a distal pointed projection, whereas Macellicephaloides improvisa has 18 segments and short cirrophores without a distal pointed projection. Macellicephaloides omitted. E, inner view of dorsal side of jaws from dissected pharynx, papillae omitted. F, inner view of ventral side of jaws from dissected pharynx, papillae omitted. G, right parapodia, posterior view (ventral cirri lost), segment 6. H, neurochaeta with spines along both margins, frontal view, segment 6. I, neurochaeta with spines along both margins, frontolateral view, segment 6. J, scanning electron micrograph of neurochaeta; up side is the distal part of chaeta. Abbreviations: fil, frontal filament; ft, frontal tubercles; inf, inflated truncate structure; pp, pointed projection. moustachu sp. nov. is a relatively common species in the CCFZ, with > 38 specimens sampled in four out of the five areas. Unfortunately, only the 18S gene could be sequenced; DNA amplification of the 16S and COI genes was unsuccessful. Consequently, identification of damaged specimens was challenging.
Etymology: The species name came from the French word 'moustachu' meaning 'with a moustache'. It refers to the palps directed ventrally, giving the impression that the worm has a moustache.
Genetic data: DNA sequencing for this species was successful only for 18S but not for COI or 16S, sharing 100% of genetic material between the specimens.
Distribution: Based on the material examined (ten specimens), this species has a wide distribution within Clarion-Clipperton Fracture Zone, being sampled in BGR, IOM (type locality), GSR and Ifremer license areas.
Reproductive notes: Paratype MNHN-IA-TYPE 1833 seems to be brooding dorsally; two large interconnected pouches were observed linked to the body through an anterior small hole on the notopodia of segment 10. The sac directly connected to the body is surrounded by a membrane and has visible small ova inside; the second sac shows an agglomerate of small ova without a visible membrane. nu gen. nov.
Remarks: Among members of the Anantennata clade (clade b3a in Fig. 2; clade d3 in Fig. 3), the genus Nu gen. nov. is unique in showing very reduced notopodia without notochaeta. Furthermore, Nu gen. nov. can be separated easily from Bathyedithia by the absence of large palpophores and from Polaruschakov by the presence of very small elytrophores and very reduced notoacicular lobe. The absence of notochaetae in subbiramous parapodia, as in Lepidasthenia and Macellicephaloides, is recognized as a generic character among polynoids (Pettibone, 1976;Barnich & Fiege, 2003) and supports the erection of the new genus.
Etymology: In the ancient Egyptian religion, 'Nu' refers to the deification of the primordial watery abyss whence all life came, also known as 'the Father of the Gods' and 'the Eldest'. Description (based on holotype): Holotype complete, 4.82 mm long and 0.76 mm wide for 18 segments (including tentacular segment), dorsoventrally flattened; slightly tapering posteriorly; live specimen translucent, bluish; digestive system and chaetae internally visible, chaetae golden (Fig. 17A); ethanolpreserved specimen pale white, translucent.
Segments 5 and 6 with large, swollen dorsal structure (Fig. 17A, B), interiorly whitish; smaller, bilobed on segment 5 (maybe because of elytrophore), larger on segment 6. Elytrophore not visible on segment 5 because of swelling. Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (longer than tip of neuroacicular lobe); in subsequent segments inserted subdistally on neuropodia (Fig. 17D), styles short (shorter than tip of neuroacicular lobe); on segments 3 and 4 shorter than those in posterior body; on segment 5 longer (approaching tip of neuroacicular lobe).
Remarks: As detailed in the genus section, Nu aakhu gen. nov., sp. nov. is unique in not having notochaetae and in having very short notoacicular lobes, very small elytrophores and neurochaetae with prominent spines along both margins.
Etymology: Again, in the ancient Egyptian religion, 'áakhu' is one of the elements that compose the human soul. An 'áakhu' is the glorified spirit or a blessed soul which has passed the final judgement (the Weighing of the Heart). The term refers to the translucent character of the body of this worm.
Genetic data: DNA sequencing was successful only for 18S but not for COI or 16S.
Distribution: Only one specimen was sampled at a single station within the Clarion-Clipperton Fracture Zone in GSR license area (type locality).
Remarks: The diagnosis of genus Polaruschakov is emended to include the following characters observed in the new species described below: jaws with small secondary tooth on margin and absence of flattened scale-like structures on segment 6. As highlighted for Hodor gen. nov., this feature, occasionally present on anterior segments, is likely to be a reproductive character, as suggested by its presence or absence without pattern among the taxa belonging to the Anantennata clade (i.e. without median and lateral antennae). polaruschakov lamellae sp. nov. tables 1,2,5) Polychaeta sp. EBS12o-Po143 (GenBank KJ736547) Janssen et al. (2015).  18A); ethanol-preserved specimen pale white; chaetae golden.
Prostomium bilobed, wider than long, lobes not developed anteriorly, short, rounded anteriorly, with an abrupt depression connecting to superior lip (Fig. 18A, B); frontal filaments absent; median notch between prostomial lobes narrow and shallow; eyes absent; a pair of internal white ganglia visible through translucent epidermis (difficult to see). Median and lateral antennae absent. Palps smooth, tapering, thin, short (reaching to segment 3; Fig. 18A, B).
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (much longer than tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia (Fig. 18D), style short (shorter than tip of neuroacicular lobe).
Morphological variation: All specimens shared the following morphological characters: short palps, chaetae, insertion and length of ventral cirri, slightly enlarged post-chaetal lobe. Although the paratypes are in poor condition, they do not seem to present the lamelliform dorsal tubercles. The tubercles might have been lost or this character could be age dependent, because the holotype has 22 segments whereas the paratypes have 18 segments.

Remarks:
The notochaetae and neurochaetae are closer to those present in Polaruschakov species. However, as described above, the lamelliform dorsal tubercles and the very reduced prostomium are unique characters, which allow differentiation of Polaruschakov lamellae sp. nov. from the other species belonging to Polaruschakov (Table 5).
Etymology: The species name came from Latin 'lamellae' meaning lamella in plural. It refers to lamelliform dorsal tubercles.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S. The specimens shared 100% of genetic material in COI and 18S, and ≥ 99.5% in 16S. The average K2P distance for intraspecific variation was 0.0% for COI and 0.2% for 16S.
Distribution: Based on the material examined (three specimens), this species has a wide distribution within the Clarion-Clipperton Fracture Zone, being sampled in BGR (type locality) and APEI#3 areas.  Prostomium bilobed, wider than long, anteriorly lobes not developed, conical; frontal filaments absent; median notch between prostomial lobes moderately wide and deep (Fig. 19A); eyes absent. Median and lateral antennae absent. Palps smooth, tapering into thin tips, short (reaching segment 3); palpophores not enlarged. Facial tubercle absent.
Segment 6 with very large (covering half of neuropodia in length), swollen dorsal structure (Fig. 19A), located between the dorsum and the base of cirrophores, interiorly whitish, of similar size.
Ventral cirri smooth, tapering into thin tips, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (about as long as tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia, style short (shorter than tip of neuroacicular lobe).
Remarks: Only two species of Polaruschakov have been described until now: Polaruschakov polaris from the Polar Basin and north of the Canadian arctic islands (730-2245 m depth); and Polaruschakov reyssi Pettibone, 1976 from the Mediterranean Sea (750 m depth). Polaruschakov reyssi is the only species having notochaetae with capillary tips. Polaruschakov limaae sp. nov. is very close to Polaruschakov omnesae sp. nov. and Polaruschakov polaris in having notochaetae with blunt tips (Table 5). However, in Polaruschakov limaae sp. nov. there are two types of neurochaetae having abrupt pointed tips and blunt tips, whereas Polaruschakov omnesae sp. nov. has one type of neurochaetae with pointed tips, and Polaruschakov polaris has one type of neurochaetae with rounded tips. Moreover, the the average K2P distance among Polaruschakov lamellae sp. nov., Polaruschakov limaae sp. nov. and Polaruschakov omnesae sp. nov. was high (24.4-24.8% for 16S). The pointed structure on the last segment resembles a minute keel, but it is not clear whether this is an artefact of fixation or a natural structure.
Etymology: This species is dedicated to Dr Maria Lima for her friendship and for initiating P.B. into the systematics of polychaetes.
Genetic data: DNA sequencing for this species was successful for 16S and 18S but not for COI.
Distribution: Only one specimen was sampled at a single station within the Clarion-Clipperton Fracture Zone in Ifremer license area (type locality). polaruschakov omnesae sp. nov.

Description (based on holotype and paratype):
Holotype complete, 4.43 mm long and 0.83 mm wide for 19 segments (including tentacular segment), dorsoventrally flattened, posteriorly tapering; colour of live animal not known; ethanol-preserved specimen pale white, slightly translucent.
Tentacular segment fused to prostomium, with a pair of short lobes, inserted laterally and slightly below prostomium; without acicula or chaetae; tentaculophores distinct, bulbous, equal sized; styles smooth, tapering into thin tips, short (reaching segment 3), dorsal tentacular style slightly shorter than ventral tentacular style (Fig. 20A); ventrally to the tentaculophores is a distinct globular pad, located laterally to the mouth. Pharynx not everted in holotype; dissected in paratype (MNHN-IA-TYPE 1842), with seven pairs of distal papillae, subtriangular, equal sized; two pairs of jaws, each one with one main fang, outer margin with a very small, secondary tooth (small elevation; Fig. 20B). Second segment with elytrophores, subbiramous parapodia, chaetae and ventral cirri.
Ventral cirri smooth, tapering into thin tips, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style short (shorter than tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia (Fig. 20C), style short (shorter than tip of neuroacicular lobe).
Morphological variation: Specimens with 19 and 20 segments were found, which share most of the morphological characters given in the species description. However, the holotype shows a minute prostomial peak and a minute ventral papilla on the pygidium, whereas the paratype does not show those peaks but presents a more rounded, small, ventral papilla.
Remarks: Polaruschakov omnesae sp. nov. is more similar to Polaruschakov polaris, with both having notochaetae with blunt tips, a wide notch and fewer chaetae than Polaruschakov reyssi. However, in Polaruschakov omnesae sp. nov. the neurochaetae tips are pointed and the palps are short (reaching segment 3), whereas in Polaruschakov polaris the neurochaetae tips are rounded and the palps are longer ( Table 5). The average K2P distance among Polaruschakov lamellae sp. nov. and Polaruschakov omnesae sp. nov. was high (23.3% for COI and 24.4% for 16S). The presence of minute prostomial peaks could be an artefact of preservation, because its presence cannot be observed in the paratype.
Etymology: This species is dedicated to Emmanuelle Omnes (Ifremer) for her help with laboratory work.
Genetic data: DNA sequencing for this species was successful for COI, 16S and 18S. Both specimens shared 100% of genetic material in COI and 16S. 18S was not successfully sequenced for the paratype. The average K2P distance for intraspecific variation was 0.0% for both COI and 16S.
Distribution: Based on the material examined (two specimens), this species has a restricted distribution within the Clarion-Clipperton Fracture Zone, being sampled in IOM (type locality) and GSR license areas. yodanoe gen. nov.
Remarks: Among polynoids of the Macellicephalinae subfamily, the presence of a cirriform dorsal tubercle can be observed in the following genera: Bathyfauvelia, Bathycatalina Pettibone, 1976, Bathybahamas Pettibone, 1985dand Vampiropolynoe Marcus & Hourdez, 2002. However, these genera differ by the number of segments, the development of notopodia relative to neuropodia and the prostomium (Pettibone 1976(Pettibone , 1985d. Yodanoe gen. nov. presents many similarities with Bathyfauvelia, such as a reduced number of segments, nine pairs of pharyngeal papillae, and notopodia shorter than neuropodia. However, the new genus presents genetic and morphological differences (Fig. 2). The K2P distance between these two genera was high (29.5% for COI and 37.8% for 16S). Furthermore, Yodanoe gen. nov. shows a triangular and tapering prostomial shape, only one type of notochaetae, an achaetous tentacular segment and jaws with a small secondary tooth, whereas Bathyfauvelia shows a short conical prostomial shape, two types of notochaetae, a tentacular segment with chaetae and serrated jaws.
Etymology: This genus is dedicated to Yoda, the Grand Master of the Jedi Order. The name is composed by Yoda and 'noe' from Polynoe, the ancient Greek nymph. yodanoe desBruyeresi sp. nov.
Prostomium bilobed, about as wide as long, lobes pronounced, anteriorly tapering to triangular peaks, located close to the notch; frontal filaments absent; median notch between prostomial lobes wide and moderately deep (Fig. 21B); eyes absent. Median antenna present, lateral antennae absent; ceratophore of median antenna, bulbous, large, short (shorter than anterior margin of prostomial lobes), inserted anteromedially on prostomium, in the notch, style missing. Palps smooth, tapering, long (reaching to segment 5; Fig. 21B). Facial tubercles absent; upper lip with few folds.
Ventral cirri smooth, tapering, present from segment 2 to last segment; inserted basally on neuropodia of segment 2, style long (reaching tip of neuroacicular lobe); in subsequent segments inserted medially on neuropodia, style short (shorter than tip of neuroacicular lobe). Parapodia subbiramous; notopodia shorter than neuropodia. Notopodia arising from dorsum as two thickened ridges; notopodia narrow, subtriangular, tapering into long acicular lobe, tip of notoacicula not penetrating epidermis. Neuropodia large, rectangular to subtriangular, tapering into long acicular lobe, tip of neuroacicula not penetrating epidermis. Notochaetae short to long, slender, slightly curved, with developed spinous rows on convex side, with blunt tips; notochaetae as stout as neurochaetae. Neurochaetae distally flattened to concave, serrated along both margins, with pointed tips.

Remarks:
The present specimen is in too poor condition to be described as a new species, but it was used to describe the pharynx for this genus. See Remarks on Yodanoe desbruyeresi gen. nov., sp. nov. for more details.
Genetic data: DNA sequencing for this species was successful for COI and 16S but not for 18S.
Distribution: Only one specimen was sampled at a single station within the Clarion-Clipperton Fracture Zone in APEI#3 area.

Bathylevensteina pettibone, 1976
Ty p e s p e c i e s : B a t h y l e v e n s t e i n a b i c o r n i s (Levenstein, 1962). The genus is monotypic. The type species was originally described as Macellicephala by Levenstein (1962), who described a particular shape of the prostomium, ending with 'bifurcate frontal horns'. Pettibone (1976) incorrectly considered these frontal horns as lateral antennae inserted medially. Later, Pettibone (1994) corrected this interpretation and placed Bathylevensteina within the subfamily Macellicephalinae.

T y p e s p e c i e s : B a t h y m a c e l l a u s c h a k o v i
The genus is monotypic. The type species was originally described and placed close to Macellicephala, Macellicephaloides and Macelloides as having no lateral antennae (Averincev, 1972;Hartmann-Schröder, 1974). Later, Pettibone (1976) revised the genus considering that the lateral prostomial horns described by Averincev (1972) were, in fact, lateral antennae. There are some misunderstandings and confusions between frontal filaments, frontal horns, cephalic peaks or lateral antennae in descriptions of polynoids. Further investigation of polynoid anatomy is required to gain a better understanding of these structures and their differences. In the present study, this species has been coded as having a median antenna and no lateral antennae.

Bathymiranda levenstein, 1981
Ty p e s p e c i e s : B a t h y m i r a n d a m i c r o c e p h a l a Levenstein, 1981. The genus is monotypic. Although Levenstein (1981) considered the presence of a median antenna uncertain, Pettibone (1985dPettibone ( , 1989c classified it in the subfamily Polaruschakovinae. We followed Pettibone and coded the species as lacking a median antenna.

Bathynotalia levenstein, 1982a
Type species: Bathynotalia perplexa Levenstein, 1982a. The genus is monotypic. Bathynotalia perplexa was originally classified as Macellicephalinae, but Pettibone (1985cPettibone ( , d, 1989c referred to it as Polynoinae based on the figure of the prostomium having lateral antennae. We followed the interpretation of Pettibone because the description in Russian was not available. However, the specimens should be re-examined to confirm the classification.

Gesiella pettibone, 1976
T y p e s p e c i e s : G e s i e l l a j a m e e n s i s (Hartmann-Schröder, 1974).
The genus is monotypic. The type species was originally described as Macellicephala by Hartmann-Schröder (1974). Pettibone (1976) incorrectly considered the presence of small spherical lobes with distal filaments as lateral antennae. Later, Muir (1982) erected the subfamily Gesiellinae Muir, 1982 for this species based on the unique diagnostic character, presence of filamentous sensory organs on the cirrophores of the dorsal cirri. However, Pettibone (1994) did not agree and finally placed this genus within the subfamily Macellicephalinae.
sinantenna hartMann-sChröDer, 1974 Ty p e s p e c i e s : M a c e l l i c e p h a l a ( S i n a n t e n n a ) macrophthalma Fauvel, 1913. Fauvel (1913, 1914b erected Macellicephala (?) macrophthalma for a single specimen in poor condition having enlarged eyes, lacking lateral antennae, and with parapodia, jaws and neurochaetae similar to Macellicephala species. According to Fauvel, the prostomium was too damaged to ascertain the presence of a median antenna. Later, Hartmann-Schröder (1974) erected a new subgenus for species lacking a median antenna. She designated Macellicephala (Sinantenna) macrophthalma as type species and included two additional species: Macellicephala (Sinantenna) arctica Knox, 1959 and Macellicephala (Sinantenna) paucidentata Eliason, 1962. Pettibone (1976 did not have access to the holotype of Macellicephala (Sinantenna) macrophthalma and considered it as doubtful, despite all the similarities highlighted by Fauvel. We believe the specimen should be re-examined in order to confirm Fauvel's observations, and species validity. The other two species that once belonged to Sinantenna were assigned to new genera by Pettibone (1976): Macellicephala paucidentata appearing to have ceratophores on the median and lateral antennae, is no longer considered as a member of Macellicephalinae; and Macellicephala (Sinantenna) arctica Knox, 1959 andMacellicephala (Sinantenna) arctica Hartmann-Schröder, 1974 (part) were synonymized with Polaruschakov polaris; and Macellicephala (Sinantenna) arctica Hartmann-Schröder, 1974 (part) was re-described as Polaruschakov reyssi.

DISCUSSION
Historically, a large number of subfamilies has been erected within the family Polynoidae in order to accommodate the great morphological diversity among genera and species. Likewise, several genera have been erected for unique species, particularly from the deep sea. Based on molecular phylogenetic analysis and morphological observations, the present study stresses that the number of subfamilies cannot be justified because many of them are not monophyletic and they include very few genera. Nevertheless, much of deep-sea diversity remains undiscovered, therefore justifying the high number of monotypic genera. The 17 newly described species, with four new genera and 138 DNA sequences based on COI, 16S or 18S, represent a significant addition to the diversity of this deep-sea group. In total, 278 polynoids were sampled during the SO239 cruise; preliminary results classified them into 44 morphotypes, but they were subsequently split into ~80 MOTUs (Bonifácio et al., 2016). These large and mobile polychaetes are best sampled with an epibenthic sledge, because the area sampled by quantitative box-corers is too small [e.g. only one polynoid was reported by De Smet et al. (2017) from box-core samples collected in the CCFZ]. Using molecular criteria for the discrimination of species almost doubled the taxonomic richness of polynoids recorded compared with morphological criteria, suggesting a high level of cryptic species diversity within deep-sea polynoids. Subtle morphological variation, however, may still allow discrimination of cryptic species, for example between Bathyfauvelia glacigena sp. nov. and Bathyfauvelia ignigena sp. nov. Interestingly, these two species are sympatric, which raises questions about the mechanisms of speciation at abyssal depths.
Unfortunately, many described genera/species fitting in Macellicephalinae sensu  were not observed in the present study or included in the molecular analysis (i.e. only 32 species in 20 genera for which DNA sequences are available). However, 89 species fitting the concept of Macellicephalinae were coded morphologically. Taken together, the representatives of 65 (molecular data) and 127 (morphological data) polynoid species from shallow to deep water allowed us to infer their phylogenetic relationships.

First steps in the Deep sea For polynoiDs
The polynoids present in our phylogenetic trees (Figs 2, 3) are clustered into two main groups: with lateral antennae and without lateral antennae. In the phylogenetic tree built from only molecular and combined with morphological data (MDS and CDS; Fig. 2), the polynoids with lateral antennae in clades a1 (mostly Lepidonotinae) and a3 are predominantly composed of shallow-water species, with only Polyeunoa laevis (a facultative commensal) reaching ~1361 m depth (Serpetti et al., 2017). This well-supported shallowwater grouping of species with lateral antennae (clade a1 mostly with Lepidonotinae members and clade a3 mostly with Polynoinae members) agrees with Norlinder et al. (2012) and is partly supported by the morphological analysis. The clade a2, composed of Bathymoorea lucasi sp. nov. and Eulagisca gigantea, which both have lateral antennae, was supported by MDS, CDS and morphological analyses. Among the subfamily Eulagiscinae, the genus Bathymoorea is the only known genus exclusively found in bathyal to The second main clade (clade b2 in Fig. 2) from the MDS and CDS analyses, which is well supported by BPP but not by MLB, included the genera Abyssarya gen. nov. and Macellicephala. The morphological analysis does not support this clade. In fact, Abyssarya gen. nov. presents two plesiomorphic characters: presence of neurochaetae with serrulations/spinous rows (character 62: 1) and absence of crenulate/ serrated neurochaetae (character 59: 0) in addition to the presence of three homoplastic characters not observed in Macellicephala genus: palps ventrally directed ( character 5: 1), neurochaetal spines ( character 55: 1) and neurochaeta distally recurved (character 56: 1). The genus Macellicephala, however, is not a monophyletic group. For instance, the type of notochaeta is not homogeneous within the genus: Macellicephala mirabilis (type species) have smooth notochaetae (character 51: 0); Macellicephala galapagensis, Macellicephala remigata, Macellicephala violacea, Macellicephala clarionensis sp. nov., Macellicephala parvafauces sp. nov., Macellicephala australis, Macellicephala laubieri and Macellicephala longipalpata have notochaetae with serrulations/ spinous rows (character 51: 1); and Macellicephala aciculata and Macellicephala atlantica have notochaetae with spines along only one side (character 52: 1). Given that notochaetae with serrulations/spinous rows are observed in polynoids with lateral antennae, this character is likely to be plesiomorphic, whereas notochaetae with spines along only one side a derived character. This suggests the need for a complete revision within the genus, which is outside the scope of the present study.
The third clade from the MDS and CDS analyses (clade b3 in Fig. 2) is represented by the genera Bruunilla, Hodor gen. nov., Nu gen. nov., Polaruschakov and Bathyedithia. This clade is partly supported by the morphological analysis. Indeed, the results show that all members without median and lateral antennae (Hodor gen. nov., Nu gen. nov., Polaruschakov and Bathyedithia) are grouped into a well-supported clade in both analyses. This clade, called Anantennata (clade b3a in Fig. 2; clade d3 in Fig. 3), showed maximal support from Bayesian analysis (BPP = 1), low support from maximum likelihood (MLB = 53-73), and was supported by the unique synapomorphy, median antenna absent (character 7: 0). Pettibone (1976) erected two subfamilies for species lacking appendages on the prostomium: Bathyedithinae and Polaruschakovinae. The main differences between those subfamilies were that Bathyedithinae has large palpophores (character 4: 1), nine pharyngeal papillae (not coded) and serrated jaws (not coded), whereas Polaruschakovinae has reduced palpophores, seven pharyngeal papillae and smooth jaws. However, Hodor gen. nov. has characters of both subfamilies, presenting large palpophores, seven pharyngeal papillae and serrated jaws, which provides support for synonymizing the two subfamilies. According to Pettibone (1976), the genus Polaruschakov has a pair of scale-like structures on segment 6 as a generic character (not coded), absent in Polaruschakov limaae sp. nov. and Polaruschakov omnesae sp. nov. as opposed to present in Polaruschakov lamellae sp. nov., suggesting that it is either a temporary character (maybe reproductive) or a specific character. Likewise, dorsal modifications between segments 5 and 8 (not coded) were observed in species belonging to the clade Anantennata without any clear pattern. For instance, the holotype of Hodor anduril gen. nov., sp. nov. showed swollen modifications on segments 6 and 8, whereas in the paratype no such modification was observed. More investigation is needed to evaluate whether these modifications are ontogenetic or reproductive and their patterns in the species. The MDS and CDS analyses suggest that the genus Bruunilla is a sister group of Anantennata, which seems consistent with the morphology considering the homoplastic characters, presence of spinous pockets or spines on notochaetae along both sides (character 53: 1), that species of the Bruunilla share with Bathyedithia berkeleyi (Anantennata) and Bathyvitiazia pallida; in addition to the presence of neuropodial papillae (character 36: 1) shared with Diplaconotum paucidendatum (Anantennata).
In summary, the present study presents the largest dataset of molecular and morphological descriptors concerning deep-sea polynoids since Pettibone (1976) and is the first major contribution about abyssal polynoids from the Clarion-Clipperton Fracture Zone. Pettibone (1976) erected many genera to account for the diversity that she observed in 37 species around the world. We believe that Pettibone has established solid morphological bases with which to increase the knowledge about deep-sea polynoids. However, the present study points out a number of morphological characters useful for the description of genera and species. A re-examination of type specimens is needed in order to progress the phylogeny of Macellicephalinae. The association of morphological and molecular analyses allows new insights into the diversity and phylogeny of deep-sea polynoids. With our 17 newly described species, deep-sea polynoids now account for 95 species in the Macellicephalinae sensu . The diversity of deep-sea polynoids still remains at one-tenth of the diversity of their shallow-water counterparts, which may suggest either that a vast reservoir of undiscovered diversity in the deep sea is awaiting discovery or that deep-sea polynoids are relatively less diverse than shallow polynoids because of lower habitat diversity or relatively recent colonization of deep waters.