New Species of Melinna (Melinnidae, Annelida) from the Australian Abyss with Comments on M. albicincta , M. cristata and M. elisabethae

. A new species of Melinna (family Melinnidae) is described from the abyssal depths off the east coast of Australia. All material was collected during the RV Investigator voyage “Sampling the abyss” (IN2017_V03) in May–June 2017 from Bass Strait, Tasmania/Victoria using a beam trawl at 4197–4133 m depth. The new species Melinna hamulus sp. nov. differs from all other species of Melinna by having around 45 abdominal segments, dorsal membrane with 13–17 projections, four pairs of branchiae completely free and occurring in depths of over 4000 m. Phylogenetic relationships between our new species and other species within the family Melinnidae were assessed using the nuclear 18S and the mitochondrial 16S and cytochrome oxidase subunit I (COI) gene fragments. The results revealed that Melinna hamulus sp. nov. was genetically distinct from all other species of Melinnidae. This is the first species of Melinna to be described from Australian waters. We provide a new standard for description of species of Melinna and provide further detail on three existing species: Melinna albicincta Mackie & Pleijel, 1995, Melinna elisabethae McIntosh , 1914 and the type species of the genus Melinna cristata (Sars, 1851) using morphological characters illustrated with light and scanning electron microscopy photographs. We provide a table with the main diagnostic characters of all described species of Melinna together with type localities and depths.


Introduction
Traditionally, Melinninae Chamberlin, 1919 and Ampharet inae Malmgren, 1866 were treated as two subfamilies within the family Ampharetidae Malmgren, 1866 andEbbe &Purschke (2021) who provided a brief description of the family Ampharetidae accepted these two subfamilies. However, Stiller et al. (2020) undertook a phylogenetic analysis of a selected number of Terebelliformia using both molecular and morphological data. They found that the subfamily Melinninae was sister to the Terebellidae Johnston, 1846 and not closely related to other ampharetids. Thus, they raised Melinninae to a new family, Melinnidae. This new family includes all the five genera previously included in the subfamily Melinninae, although molecular data for Melinninae used in Stiller et al. (2020) was limited. The authors used molecular data from 41 taxa of Ampharetinae and only six taxa of Melinninae. All other genera previously classified in the Ampharetinae have been elevated to the Ampharetidae and, within this, Stiller et al. (2020) recognized three clades: Ampharetinae, Amaginae, and Amphicteinae, although it is unclear if these should be considered as subfamilies, given the limited numbers of species in this family sequenced.
Melinna are tubicolous annelids and members of the genus share the presence of strongly recurved dorsal hooks behind their branchiae. The resemblance of these curved hooks to grappling hooks gives them their common name "grapple worms" (O'Reilly & Nowacki, 2008). Members of the genus have been recorded in very high densities, for example, in Galway Bay, Ireland animals were found in densities of around 1500 individuals/m 2 (Grehan et al., 1991) and even 5000 individuals/m 2 off the Northumberland coast, UK (Buchanan, 1963;Hutchings, 1973) where they are an important food source for cod. Furthermore, their high densities make them key sediment stabilizers of intertidal mudflats as seen in the Arcachon Bay, France (Bernard et al., 2014).
Here we report one new species belonging to the genus Melinna from deep waters along the Australian eastern continental margin. The phylogenetic position of the new species within Melinnidae is assessed using molecular data.

Study area
Samples were collected during the research vessel RV Investigator voyage "Sampling the Abyss" (IN2017_V03) to deepwaters (900-4,800 m depth) off the east Australian coast (O'Hara, 2017) using a CSIRO 4 m wide × 0.5 m high Beam Trawl (Lewis, 2010). From 15 May to 16 June 2017 samples were taken along a south to north latitudinal transect of 18 degrees along the east coast of Australia, from 42° to 24°S (Fig. 1, showing type locality relative to the Australian coast) at target depths 2,500 m and 4,000 m. Specimens of Melinna were collected during operation (OPS 30) from Bass Strait, Tasmania/Victoria at 4,197-4,133 m. Onboard specimens were live sorted into higher taxa on ice in chilled (5°C) seawater, annelids were sorted to families, then fixed in either 95% ethanol or in 10% buffered formalin. At the Australian Museum Research Institute laboratory, formalin fixed specimens were soaked in water and preserved in 80% ethanol. All material was then sorted to genera.

Morphological analysis
Comparative material of Melinna cristata and M. elisabethae was examined from the Australian Museum collections. Melinna albicincta was loaned from the National Museum of Wales (NMW).
The specimens were examined in ethanol using a dissecting microscope (OLYMPUS SZX7) and compound microscope (OLYMPUS BX53). Specimens were extracted from their tubes, measured (length from prostomium to pygidium), stained with methyl blue and photographed using an OLYMPUS DP74 camera attached to dissecting and compound microscopes with the imaging software OLYMPUS cellSens Standard 1.17.
One of the paratypes of Melinna hamulus sp. nov. and material listed in Mackie & Pleijel (1995) including a paratype of M. albicincta and other material of M. cristata and M. elisabethae was dehydrated in ethanol, critical point dried, coated with 20 nm of gold, and examined under the Scanning Electron Microscope (SEM) JEOL JSM6480 at Macquarie University, Sydney. All material of the new species is lodged at the Australian Museum, Sydney (AM); all AM registered specimens are prefixed "AM W". Material fixed in 95% ethanol was used for molecular analysis.

Sequence analysis
Overlapping fragments were assembled into consensus sequences and edited in Geneious Prime 2019.0.4 (https:// www.geneious.com). A BLAST analysis (Altschul et al., 1990) was performed to confirm the correct region had been amplified, to compare with other sequences on GenBank, and to check for contamination. New sequences were submitted to GenBank (Table 1). Additional sequences from the family Melinnidae (10 taxa) were downloaded from GenBank (Table 1). Two species, Polycirrus carolinensis and Terebella lapidaria, were used as an outgroup.
Sequences were aligned using the Geneious plugins with the default settings: MAFFT (Katoh et al., 2002) for 16S and 18S and MUSCLE (Edgar, 2004) for COI. Pairwise genetic distances for 16S and COI were calculated for aligned sequences in Geneious. Concatenated sequences for all three genes were made in Geneious. JModelTest (Darriba et al., 2012) was used to find the best model using the Akaike information criteria. The model GTR + I + G was selected as the best model for each gene. Phylogenetic trees were constructed in IQTREE. Trees were visualized in FigTree v1.4.4 (Rambaut, 2018) and edited in Adobe Illustrator.

Molecular results
Genomic DNA was successfully amplified and sequenced from three specimens of Melinna hamulus sp. nov. (AM W.50732, AM W.50353 and AM W.50382). The combined data of three concatenated markers had 2918 aligned positions (COI with 654 positions, 18S rDNA with 1778 positions and 16S rDNA with 486 positions).
Maximum Likelihood analysis of combined COI, 16S and 18S sequence data formed two major clades within Melinnidae (Fig. 2)

Remarks
The family Melinnidae was originally a subfamily within the family Ampharetidae Chamberlin, 1919. Day (1964 emended the diagnosis of the subfamily and provided two tables listing the characters of genera and type species in the subfamily. As Day (1964) states, many new genera have been described since Chamberlin erected the subfamily. Day (1964) emended the diagnosis to include all members that have acicular neurochaetae on segments 2-5 (Day incorrectly numbers them as present on segments 3-6) and indicated that members of this subfamily may or may not have dorsal hooks and paleae are always absent. Later, as a result of a molecular and morphological study of the Terebelliformia, Stiller et al. (2020) elevated this subfamily to a full family, although they did not formally provide a diagnosis of the new family. Subsequently Rouse et al. (2022) provided a description of the family, including morphology, physiology, natural history and systematics, but no diagnosis. The above family diagnosis is based on Stiller et al. (2020) and Rouse et al. (2022) although there are comments in Rouse et al. (2022) such as "some species show unusual neurochaetae on anterior segments", thus it is unclear if these characters are restricted to particular genera or not. Furthermore, Rouse et al. (2022) clearly state that all Melinnidae have straight acicular spinelike neurochaetae on 2nd to 4th or 5th segments (chaetigers 1-4) which is contradictory to the previous statement. Similarly, Rouse et al. (2022) suggest that notopodia can begin on segment 4, although usually they begin on segments 5 or 6 and it is unclear if this is a generic character or not. Rouse et al. (2022) neither indicate the segments on which the dorsal hooks occur nor the genera in which they occur. According to Holthe (1986), these hooks are present on segments 3 and 4, but Rouse et al. (2022) do not agree with the segmental numbering system of Holthe (1986) who did not recognize the achaetous segment 1, thus all Holthe's numbering is one number out. We suggest that all the genera in this newly erected family need to be rechecked to clarify the above characters, but this is beyond the scope of this paper. Type species Sabellides cristata Sars, 1851 by monotypy.

Remarks.
The above diagnosis is based on Mackie & Pleijel (1995), but is modified with regards to segment number, as Rouse et al. (2022) state that there is an achaetous segment (segment 1) before the first chaetiger (segment 2). While the branchiae arise from segments 2-4, they appear to arise from a dorsal ridge at the level of segments 2-3, and they exhibit varying levels of basal fusion. Nephridia are on segments 4-7. With regard to the segment on which the dorsal hooks appear we do not follow Holthe (1986), instead we follow Stiller et al. (2020) and Rouse et al. (2022) to indicate that they begin on segment 3. Earlier descriptions of Melinna species can be quite confusing regarding the numbering of segments and chaetigers. Most authors of previous descriptions have failed to recognize or neglected the anterior achaetous segment behind the peristomium, while the anterior notopodia are small and ventrally displaced so may have been missed. While the definition of Mackie & Pleijel (1995) indicates that notopodia begin from segment 4 or 5, we cannot find any records of species with notopodia from segment 5, and the confusion in the literature about anterior segment numbering does not help. Table 3 lists the numbering of anterior segments which will facilitate the interpretation of earlier descriptions of species of Melinna as many lack the presence of an anterior achaetous segment.
We have also followed more recent papers in referring to segments by numbers and not by Roman numerals. Given the confusion of some species from northern Europe, including the type species of the genus Melinna, M. cristata, we have provided expanded descriptions and figures to supplement those provided by Mackie & Pleijel (1995

Description
Description based on holotype with variation observed among paratypes indicated. Prostomium with welldefined anterior and posterior parts, separated by pair of transverse nuchal slits (Fig. 3F), anterior part distally trilobed, followed by achaetous segment forming a glandular collar dorsally with smooth margin (Fig. 3B,F). Buccal tentacles almost completely withdrawn with tips exposed (Figs 3F, 4A,B). Lateral wings of anterior body between prostomium and segment 5 highly arched (Fig. 3A,B,C). First four segments glandular, subsequent seven segments with discrete ventral glandular shields (Figs 3C, 4A,C). Four pairs of smooth branchiae on segments 2-5, although appearing to arise on a dorsal ridge on segments 2 and 3, arranged two in front and two behind. Completely separated from each other. All branchia similar in both width and length, tapering to fine tips distally and having slightly swollen bases with slightly crenulated surfaces (Fig. 4F,G,I). Branchia circular in cross section. Incomplete dorsal membranous fold on segment 4 (Fig. 3B,C). Postbranchial dorsal membrane with 13 (17) triangular lobes or projections, all similar in size, except for the slightly larger middle one, some bilobed, and middle larger one with small lateral lobes at base so appearing as tricorn ( 3D). Neurochaetae as an elongated row of simple acicular short spines on slightly raised glandular ridge and inserted progressively more dorsally from segment 2-5 (Fig. 3A,B). Neuropodia with pectinate uncini from segment 6, present in 21 pairs of thoracic uncinigers (to end of abdomen, segment 26). Numbers of pairs of abdominal neuropodia 45, based on AM W.53257. Abdominal neuropodia elongate and basally swollen, without any rudimentary notopodia (Fig. 5E,G). Thoracic uncini from segment 6 with 30 uncini within a row.
Abdominal neuropodia with 31 uncini in torus. Dentition of uncini within a row varies slightly (Fig. 5F,H). Thoracic uncini avicular with two teeth in a vertical row over rostral tooth, subrostral process and basal prow (Figs 3G, 5A,B,C). Side of uncini with marked ornamentation (Fig. 5C). Abdominal uncini with conspicuous long tendons attaching them to body and visible through body wall. Uncini arranged on elevated neuropodia in curved row (Fig. 5E,G). Uncini with four to six teeth (Fig. 5F,H). Posterior segments compacted, none of the paratypes complete. Pygidium with two lateral lobes and 4-5 smaller lobes surrounding them (Fig. 3H). No anal cirri present.
Tube. Composed of finegrained sediment and lined with a thin, stiff clear membrane. Length of tube at least 2 times specimen (Fig. 3E).
Methyl blue staining pattern. Use of methyl blue staining reveals strong staining of the prostomium and segments 1-4, branchiae lightly speckled at tips and transverse rings at bases. Stained band behind dorsal membrane from segments 5-6. Light staining elsewhere on thorax and on abdomen. Light staining on ventral shields, staining stronger laterally on shields.

Remarks
Melinna hamulus sp. nov. is characterized by the following characters: around 45 abdominal segments, dorsal membrane with 13-17 triangular projections, four pairs of branchiae completely free, one pair of curved dorsal hooks with pointed tips lacking a canal, and a pygidium with two large lateral lobes surrounded by four or five smaller lobes, which distinguishes it from all other described species where the data are available (Table 5). We compare the morphological characters along with type locality and depth of our new species with the other currently accepted 26 species of Melinna (Tables 4, 5). Gunton et al. (2021) Tables 4, 5 Melinna albicincta Mackie & Pleijel, 1995: 116-120, fig. 3, 5,

Remarks
The original description only provided line drawings, here we give additional information based on light and SEM photographs. Prostomium with two elongate transverse nuchal slits (Fig. 6A,B). The first segment achaetous, forming a ventral collar with an elevated crenulated anterior margin (Fig. 6D). Segments 2, 3, 4 and 5 with neurochaetae as short acicular spines with neuropodia becoming more lateral, on segment 6 elongated raised tori with neurochaetae as uncini (Figs 6D,  7B). First notopodia on segment 4, very small with short capillary chaetae, and gradually becoming more dorsal and better developed by segment 6 with two tiers of capillary chaetae (as seen in chaetiger 11, Fig. 7C). This numbering of anterior segments differs from that given by Mackie & Pleijel (1995) as they did not recognize an achaetous segment 1 which forms a ventral glandular collar (Fig. 6D). Pair of short dorsal curved hooks (Fig. 6E, F). Four pairs of elongate branchiae ending in tapered pointed tips (Fig. 6A,D,E), with inner and anterior most pair basally fused (35-40%) (Figs 6E, 7E,F); two conspicuous white bands present on inner pairs (Fig. 6A,B) but absent on other branchiae. Segment 5 with dorsal membrane (Fig. 7E) bearing large number of small projections. Larger middorsal papilla beneath dorsal membrane (Figs 6E, 7E,F), as also shown in Fig. 5A of Mackie & Pleijel (1995). Ventrum with anterior margins of segments elevated and glandular (Fig. 7A). Notopodia with elevated lobes and notochaetae arranged in two tiers (Fig. 8A,D). Anterior neuropodia forming raised tori with uncini arranged in a curved row (Fig. 7B), with four teeth in longitudinal row on chaetiger 11 and 13 (Figs 7D, 8B respectively). Rudimental notopodia on the abdomen as small rounded projections (Fig. 6C). Mid and posterior neuropodia becoming increasingly elevated (Fig. 8A,F). Abdominal uncini with four teeth in a longitudinal row (Fig. 8B,C,E). Pygidium with terminal anus and surrounded by a pair of large lateral lobes and a number of smaller ones (Fig. 8F). Living in muddy tube with attached shell fragments (Fig. 6G).

Remarks
The redescription by Mackie & Pleijel (1995) only provided line drawings, here we give additional information based on light and SEM photographs.

Remarks
The original description by Mackie & Pleijel (1995) only provided line drawings, we give additional information based on light and SEM images.
Prostomium welldeveloped and divided into anterior and posterior parts, anterior part trilobed (Fig. 15B), posterior part raised triangular glandular with anterior margin bearing longitudinal slits (Fig. 15B). Buccal tentacles numerous and equal in length to the branchiae (Fig. 13B). Four pairs of branchiae annulated with distal surface exhibiting cilia patches (Fig. 14A,C). Branchiae circular in crosssection, anterior pair showing 40% of basal fusion and outer pair 20% fusion (Figs 13B,C, 14C). One pair of dorsal hooks strongly curved and with pointed tips (Figs 13B,D,E, 14B,C), solid without internal canal. Dorsal membrane with about 12 triangular projections on anterior margin (Figs 13E, 14A). First segment achaetous (Fig. 14A) with crenulated ventral margin (Fig. 15B). Segments 2-5 with acicular neurochaetae arranged in elongated tori, progressively inserted more laterally (Fig. 13B,C). From segment 6 neuropodia with uncini continuing to pygidium. Notopodia present from segment 4, initially very small with short notochaetae, then progressively becoming larger, more laterally inserted and welldeveloped by segment 7 (Figs 13D, 14A, 15C) and on all subsequent ones. Notochaetae arranged in two tiers ( Fig.  14H) with tips of longer ones curved (Fig. 15A) and with blades finely ornamented (Figs 14F,H,15A,E). Notopodia globular with posterior lobe longer than anterior one (Fig.  14D,E,F) and well separated from neuropodia (Figs 14E,15C). Thoracic neuropodia raised glandular tori slightly curved, bearing about 26 uncini (Fig. 14E). Thoracic uncini with a single row of four teeth with ornamentation on lateral face and becoming marked and behind anterior tooth strongly ornamented with small teeth (Figs 14G, 15D,F). Abdominal rudimentary notopodia not observed as in Mackie & Pleijel (1995), specimens very small and thin (Fig. 15G). Abdominal neuropodia welldeveloped but not elevated, anteriorly bearing four teeth arranged in a single row (Fig. 15H,I). Terminal anus bounded by pair of large lateral lobes and several smaller lateral ones (Fig. 13a) and lacking anal cirri. Methyl blue stain shows extensive ventral glandular areas (Fig. 13F). Tube with attached shell fragments on anterior end and rest of tube made of smaller sand grains and mud (Fig. 13G).

Discussion
To date 26 species of Melinna have been described and accepted including this new species Melinna hamulus sp. nov. This is the first species of Melinna to be described from Australian waters. We have described our material as a new species, based on the morphology and that the new species is genetically distinct from all other species of Melinna for which DNA sequence data are available.
When comparing species of Melinna using morphology, we noted that many species were poorly described and illustrated, and in some cases no figures were provided (Table 5). We used the numbering system from the original descriptions as it is unclear if the original descriptions' numbering system recognizes the presence of the first anterior achaetous segment or not. These poor original descriptions have not prevented some of these species, such as the type species of the genus M. cristata (Sars, 1851), being widely recorded from localities far away from the Norwegian type locality (see Parapar & San Martín, 1997) and for a subspecies Melinna cristata australis Hartmann Schröder, 1965 being described from Chile. Mackie & Pleijel (1995) began a revision on just the species occurring in northern Europe and found that the species complex of M. cristata included a new species M. albicincta Mackie & Pleijel (1995). These authors resurrected M. elisabethae McIntosh, 1885, a species previously synonymized with M. cristata. Mackie & Pleijel (1995) concluded that another species from Iceland, M. islandica Saemundsson, 1918, whose type is lost, must be considered as a nomen nudum. Following on from Mackie & Pleijel (1995), Parapar & San Martín (1997) described a new species M. arnaudi from the Antarctic and discussed at length how the Antarctic records of M. cristata probably do not belong to this species, while the records from South Africa by Day (1967), may represent a sibling species of their new species. We suggest that all these records from the Antarctic and subAntarctic (Ehlers, 1887(Ehlers, , 1908Monro, 1930;HartmannSchröder & Rosenfeldt, 1965 as M. cristata australis; Hartman, 1966Hartman, , 1967Hartmann Schröder & Rosenfeldt, 1989 should be reexamined.
The misidentifications, lack of detail in early descriptions and lack of genetic material from type localities all highlight the need for a revision of the genus Melinna, and indeed of the whole family Melinnidae. Type material of all species should be reexamined or fresh topotypical collected examined and sequenced. Table 5. Diagnostic characters of Melinna species. Column headings are abbreviated as follows: corrected, numbering of anterior segments corrected; hooks, numbers of pairs of dorsal hooks; fusion, numbers of branchiae and degree of fusion; projections, dorsal membrane on segments 5-numbers of projections; neurochaetae, numbers of neuropodia with acicular neurochaetae; and segments, numbers of abdominal segments. *Numbering of segments as stated in original paper indicating lack of recognition of achaetous segments, difficult to reconcile with Rouse et al. (2022).