Contribution to the taxonomic knowledge of Ampharetidae (Annelida) from Antarctica with the description of Amage giacomobovei sp. nov

Thanks to newly collected material from the Terra Nova Bay area (Ross Sea, Antarctica), we discuss the taxonomy of the ampharetid genera Amage Malmgren, 1866 and Amythas Benham, 1921. A new species of Amage, A. giacomobovei sp. nov., is described based on morpho-anatomical data. This is the second new species described from an area which appears to be rich in ampharetids, a coastal embayment at ~500 m depth near the Italian “Mario Zucchelli” research station. The new species is characterized by having 16 abdominal uncinigers and four pairs of branchiae that readily distinguish it from its congeners. Tubes of A. giacomobovei sp. nov. are also characteristic in showing a large amount of embedded sponge spicules, suggesting a possible close association to spicule mats. Based on the amended diagnoses of the two genera, Amage septemdecima Schüller & Jirkov, 2013 is transferred to the genus Amythas. Finally, to simplify the task of ampharetid genera recognition for untrained people, we provide a dichotomic key for ampharetid genera found in Antarctica and a checklist of species occurring in Terra Nova Bay.

In Antarctica this group shows very high densities and diversity (Wlodarska-Kowalczuk et al. 2007;Grange & Smith 2013) with many species only recently described (Schüller 2008;Schüller & Jirkov 2013;Schiaparelli & Jirkov 2016), suggesting that many more species are waiting to be discovered.
In this paper we discuss the taxonomy of the ampharetid genera Amage Malmgren, 1866 andAmythas Benham, 1921, emend the description of its type species, A. membranifera Benham, 1921, and describe a new species of Amage. This new taxon was found in an embayment at ~500 m deep (nicknamed "the canyon"), located at the centre of the Gerlache Inlet (Terra Nova Bay), which seems to act as a natural sink for the organic matter produced during Austral summer months. From this same site we have already described another species from this family, Amphicteis teresae Schiaparelli & Jirkov, 2016. Finally, we provide a dichotomic key for ampharetid genera occurring in Antarctica, with a particular focus on those species occurring in Terra Nova Bay, part of a species inventory in the framework of research projects of the Italian National Antarctic Research Program (PNRA).

Taxon sampling
The material analysed herein were collected in the Terra Nova Bay area (Supp. file 1), off Mario Zucchelli Station (74.70000º S, 164.1166700º E), in a deep depression nicknamed "the canyon", during two different expeditions of the Italian National Antarctic Research Program (PNRA): the XXVIII th (2012)(2013) and the XXIX th (2013)(2014). Materials from these expeditions were studied in two different research projects, i.e., "BAMBi" (Barcoding of Antarctic Marine Biodiversity, PNRA 2010/ A1.10) and "ISOBIOTOX" (stable ISOtope and molecular markers for the reconstruction of Antarctic trophic webs under the sea-ice influence: evaluation of robustness to BIOdiversity loss and heavy metals bioaccumulation-neuroTOXicity, 2013-AZ1.16).
The type material, herein described, was obtained through a pelagic SHPN net (Small Hamburg Plankton Net, Hydrobios, Germany) which accidentally hit the bottom instead of remaining in the water column. The SHPN opening was of 1 m 2 , with a mesh size of 2 mm. These specimens were obtained by sorting the accidentally collected benthic material (roughly 1 m 3 ). Specimens of Amythas membranifera were collected the following year, at the same site, by the "ISOBIOTOX" project. In this case, samples were obtained by using a rectangular dredge (mouth 40 × 20 cm, with 1 cm mesh at the cod end).

Photography and laboratory analyses
Collected specimens were immediately placed in 1°C seawater and photographed alive, after having gently opened the tubes with a dissecting scissor. Photographs were taken by using a Nikon D700 reflex camera, equipped with a 105 mm lens and external, infrared-triggered, flashes (Nikon SBR200). Specimens were then fixed in 95% ethanol or frozen for molecular and isotopic analyses. Supplementary laboratory photos of fixed specimens were later taken with a Leica DFC490 camera, stereo microscope Leica M165C, and compound Leica DMI 4000B microscope at the PP Shirshov Institute of Oceanology of Russian Academy of Science, Moscow. For better contrast specimens were been stained with a solution of methylene blue (not methyl blue), which gives better results than methylene green. For scanning electron microscopy (SEM), specimens stored in 70-75% ethanol were placed in 100% ethanol and 100% acetone, then critical point dried, using СO 2 as a transition fluid. After drying, specimens were sputtered with gold for SEM observation. SEM micrographs were taken in a Camscan S-2 Cambridge instrument Scanning Electron Microscope at the M.V. Lomonosov User Facilities Centre (Moscow State University).

Diagnosis (amended)
Prostomium. Prostomium trilobed, middle lobe anteriorly incised or with horns, without longitudinal ridges, with couple of nuchal organs at posterior margin of middle lobe. Lower lip not enlarged and longitudinally grooved.
BraNchiae. 3-4 pairs, if four pairs of branchiae are present, these are arranged nearly segmentally with 2 pairs on segment 3 and 1 pair on each S4 and S5 (2+1+1). When branchiae are in three pairs only, it is the S5 lacking these.
aBdomeN. Abdominal rudimental notopodia well developed, but not enlarged. Number of AU 7-21, usually constant for species.

Remarks
The presence of nuchal organs has been mentioned for Amage by Moore (1923), Hilbig (2000), Schüller & Jirkov (2013) and Reuschert et al. (2015). Hilbig (2000) modified the diagnosis of Amage to include NO. However, none investigated the type species of Amage, i.e., Amage auricula Malmgren, 1866. The types of A. auricula are lost (Holthe 1986), but more than 2000 specimens of this species from about 200 localities from the North Polar Basin (see map in Jirkov 2001: 451), including specimens near the type locality, have been investigated. In all these specimens there are well developed NO, well recognizable in the stained specimen reported in Fig. 1B. Hilbig (2000) and Reuscher et al. (2015) included in generic diagnosis of Amage the presence of smooth buccal tentacles. In the case of the new species herein described, BT are obviously not smooth but, at the same time, this species is beyond any doubt an Amage. Therefore, we consider that the shape of BT cannot be included into the generic diagnosis of Amage. This consideration should also be extended to the definition of other ampharetid genera, as was already suggested by Jirkov (2011). Jirkov (2011) proposed Egamella Fauchald, 1972, Mexamage Fauchald, 1972, Paramage Caullery, 1944and Phyllampharete Hartman & Fauchald, 1971 as junior synonyms of Amage. Reuscher et al. (2015) accepted synonymy of the first three genera but rejected the synonymy of Phyllampharete. We agree with Reuscher et al. (2015) and hence Phyllampharete is not included in the list of Amage synonyms here either.
The taxonomic status of Amage anops perfecta Moore, 1923 is unclear. The author of the species described the status as follows: "This species, at first thought to be distinct under the name A. perfecta, is now regarded as identical with Johnson's species [A. anops] or at most as only a subspecies" (Moore 1923: 210) and a taxon never described as new species. Since then, the type material has not been reexamined.

Diagnosis
The new species is characterized by having 16 AU and four pairs of branchiae.

Etymology
This species is dedicated to the Italian piemontese explorer Giacomo Bove (1852-1877, https://en.wikipedia.org/wiki/Giacomo_Bove), lieutenant of the Italian Royal Navy, who joined the Vega expedition of Adolf Erik Nordenskiöld (1878-1879) in search of the North-East Passage. He was the first Italian to attempt to organize an expedition to Antarctica, however, it was not funded due to its prohibitive costs following the union of Italy.  Prostomium. Prostomium with T-shaped middle lobe encircled by inflated lobe (trilobed); middle lobe with eyespots at the posterior corners; couple of NO present along the posterior margin, separated by a narrow gap; without glandular ridges (Fig. 1A).

Holotype
Buccal teNtacles. Buccal tentacles few, smooth; tentacle upper surface with rows of warts stained with methylene blue much more intensively than other tentacles' surfaces ( Fig. 1C).
BraNchiae. Four pairs of branchiae, places of attachment of two branchostyles in transversal row, large median gap (equal to several diameters of branchostyles) between groups of branchiae; branchiae gap with longitudinal folds; other two branchostyles form longitudinal line behind the first two; branchostyles cirriform, smooth, without additional structures visible even after staining; anterior outermost branchiae originating from C1, next from C2 and posterior branchiae from C3; anterior innermost branchiae originating from S2.

Methylene blue staining
Anterior halves of the lower lip and ventrum of the next S2 or S2 only stained differently from their posterior halves: immediately after staining with dense small dark blue patches, later when the stain moves inside the body, patches are still dark, while the stain from the posterior halves dissolves. There is a dark violet or blue band between notopodia ventrally (except for one-two last thoracic segments) (Fig. 1E). Parapodia and ridges connecting notopodia and neuropodia (both thoracic and abdominal) ( Fig. 1G) and branchiae ( Fig. 1E) stained more or less intensively than the rest of the body. A scattered distribution of small intensively stained spots, denser in neuropodia. Spots form narrow strips behind bands connecting neuropodia (Fig. 1H).

Differential diagnosis
Only four species of Amage have 15 TC and 12 TU, but they differ by smooth buccal tentacles (not known for A. tasmanensis) and the number of AU. Amage imajimai described off Japan from a depth of 990-1060 m and A. tasmanensis described from Tasman Sea (3830 m deep) have 11 AU. Amage imajimai has thoracic uncini with two rows of teeth and abdominal uncini with several rows of teeth. Amage tasmanensis has both thoracic and abdominal uncini with a single row of teeth. Amage longitorus described off Japan from a depth of 1060-3016 m has 12 AU, and three pairs of branchiae instead 16 AU and four pairs of branchiae. Amage benhami known from a bioherm off Oregon coasts (North East  Pacific, 44.669100º, -125.111417º, depth 625 m), has 15 AU instead of 16 AU. Amage benhami also differs in the shape of thoracic (each uncinus has a single row of teeth) and abdominal uncini (each uncinus has numerous teeth arranged in three non-parallel vertical rows), while in A. giacomobovei sp. nov. both thoracic and abdominal uncini are similar.  16 AU as it was correctly reported by Reuscher et al. (2009) and thus belong to A. giacomobovei sp. nov., not to A. benhami.

Distribution
Only known from the Ross Sea (Antarctica) in 290-500 m. Benham (1927) described the tubes of A. giacomobovei sp. nov., which he considered to be A. sculpta, as filled with sponge spicules: "The tube if of grey mud with abundant sponge spicules embedded and foreign bodies adherent. The tube measures 35 mm. by 4 mm. at its upper end. The contained worm is 15 mm. in length". Although no data are available about the bottom type where our specimens and those examined by Bentham were collected, it seems probable, given the high amount of sponge spicules embedded in the tube walls, that this species might be typically occur only associated to "spicule mats" fields (Gutt et al. 2013). Amage giacomobovei sp. nov. tubes are also quite similar to those of A. auricula, type species of the genus, which occurs in the North Polar Basin and North Sea. The high level of variation in pygidium later cirri shape is probably the result of different levels of contraction of cirri due to fixation.

Diagnosis
Prostomium. Prostomium of Ampharete-like, subdivided by a U-shaped groove; median lobe anteriorly rounded; nuchal organs and ridges absent; lower lip not enlarged and longitudinally grooved; nephridial pores/papillae not visible.

Remarks
Jirkov (2011) proposed the inclusion of Amythas in Neosabellides Hessle, 1917 with certain doubts. However, the new material from Antarctica available for this study clarifies the position of Amythas, which is here considered valid, and its type species redescribed. The status of Neosabellides has been clarified earlier (Jirkov 2018).
Amythas has an Ampharete-like prostomium but it differs from all other known genera of Ampharetinae Malmgren, 1866 by the following combination of characters: i) absence of nephridial pores/papillae behind branchiae; ii) middle lobe of prostomium anteriorly rounded; iii) absence of nuchal ridges and any other structures of the prostomium; iv) thoracic and abdominal neuropodia of the same shape; v) absence of modified notopodia; vi) absence of dorsal ridges; vii) absence of enlarged neuropodia; viii) uncini of thoracic and abdominal neuropodia with similar morphology.
The genus Amythas also differs from Ampharete Malmgren, 1866 by the absence of a pair of nephridial papillae behind branchiae. In this aspect Amythas is close to Amage, and that is why Amythas septemdecima Schüller & Jirkov, 2013 was initially placed in Amage. Due to the above considerations, Amage septemdecima (Schüller & Jirkov, 2013) is here transferred to the genus Amythas due to the combination of characters such as the shape of prostomium, the absence of nephridial papillae behind the branchiae, the similarity of uncini and neuropodia throughout the body and the absence of any specific characters of thorax. This is in accordance to Reuscher et al. (2015), who disagreed with the fact that A. septemdecima could be settled in Amage. Benham, 1921 Figs 4, 5A-E, J-L Amythas membranifera Benham, 1921: 102-105, pl
BraNchiae. Three pairs of branchiae, the outermost one originating from TC1, the innermost one from TC2, and the middle one from S2 (Fig. 5C). Branchostyles cirriform, short, organized in a straight line and showing a wide middle gap. Branchostyles and branchophores covered with warts. There are no visible nephridial papillae or nephropores.
NotoPodia aNd NeuroPodia. 17 TC, notopodia with capillary chaetae from S3; anterior notopodia small, increasing in size from first to third pair; elevated or modified notopodia absent; 14 TU, all thoracic neuropodia well developed, with almost equal size, middle ones slightly bigger than anterior and posterior ones. Ventral shields distinct anteriorly and completely disappearing before the end of thorax. Subdivision of thorax absent. 17-20 AU; shape of neuropodia gradually changing throughout in most specimens, from tori to pinnula with marked size reduction on the thorax/abdomen transition; uncini at the margin of neuropodia. All neuropodia without cirri; typically rudimental abdominal notopodia absent. Pygidium. anal cirri absent.
tuBe (Fig. 4F-G). Tubes of this species appear to be formed by progressive additions of small quantities of homogeneous fine-grained sediment to form an irregular banded pattern (Fig. 4F). The interior part of the tube has a homogeneous and transparent membranous lining that isolates the body of the worm from the outer layer and that can easily be peeled off from it (Fig. 4G).

Distribution
This species was reported from fjords along the West Antarctic Peninsula (WAP) (Grange & Smith 2013) and in several stations from the Ross Sea (NZ NIWA "IPY-CAML" Voyage TAN0802; records retrieved from GBIF, last accession 2020 Feb. 20).

Remarks
Benham in the description of the new genus Amythas, stated that there were no buccal tentacles and that these were replaced by a folded membrane (Benham 1921). Hartman (1966) and Fauchald (1977) followed Benham. As it has already been stated by Monro (1939), buccal tentacles are present in Amythas as in all other Ampharetidae. The whole construction of buccal tentacles shows perfect resemblance to Terebellidae Johnston, 1846 and, at the same time, to Ampharetidae with an everted pharynx (see Jirkov 2016: fig. 3) which provides an indication of their homology and prostomial origin. The position of the uncini at the margin of the neuropodia was found to vary according to the individuals (e.g., Fig. 5D-E). The smallest specimen examined (MNA-07935) has small rudimental abdominal notopodia and AU neuropodia more pinnuli-like (Fig. 5D). The morphology of the tube of this species was documented and described to occur in extant Sabellidae as well as in the ichnofossil Caprascolex antarcticus Schweitzer et al., 2005 from the Antarctic Eocene of La Meseta formation (Seymour Island, Antarctica) (Schweitzer et al. 2005: figs 3b, 4). In situ images of tubes of this species are available in Grange & Smith (2013: fig. 2c).

Key to Ampharetidae genera occurring in Antarctica
In order to simplify the task of genera recognition of not trained people we have assembled a dichotomic key hoping that this would speed the recognition of the different Antarctic ampharetid genera.

Discussion
The availability of newly collected material from Terra Nova Bay has enabled the description of a second new species of Ampharetidae, Amage giacomobovei sp. nov, after that of Amphicteis teresae, and allowed a better understanding and definition of some morphological characters, that have been debated for a long time, of the genus Amythas.
This peculiar genus differs from all other genera of Ampharetinae in having a unique combination of characters (see Remarks above) that should make its diagnosis straightforward. On the other hand, it is exactly this mix of characters, some of which have ‛intermediate' states, which makes the evaluation of the affinities of this genus, compared to other Ampharetidae on a morphological base, difficult. These 'incongruences', for example, lead to the proposal (Imajima et al. 2012) of a new term (i.e., "intermediate uncinigers") to designate segments carrying neuropodia of tori type but avoid of notopodia and notochaetae and, in turn, to a new definition of body regions in Amythas. However, these uncertainties are here reinterpreted and, at the same time, simplified, thanks to the newly available material. Of course, natural affinities and remaining uncertainties could only be solved when molecular data is available to build a robust phylogenetical framework for the family.
One of the main differences between Amythas and other genera is represented by the shape of neuropodia and their uncini. Ampharetidae has two types of neuropodia: tori and pinnuli (Figs 1E, G and 5H-I). Tori have uncini situated in a furrow relatively far from the margin of neuropodia, while pinnuli have uncini exactly at the margin of neuropodia. Tori are bigger and more sessile than pinnuli. In most ampharetid genera anterior neuropodia are tori and posterior are pinnuli. This change is abrupt and often specific for genera.
Since Phyllocomus and Samythella show some characters, such as a prostomium not subdivided into lobes, number of AU large and variable within species, that can be considered plesiomorphic (e.g.), we suggest that the abrupt change in the two types of neuropodia having a fixed position could represent an apomorphic character. Although only a robust molecular phylogenetic framework would determine which are the true plesiomorphic and apomorphic characters, these observed changes may serve as working hypotheses to be tested with molecular data in the future. According to this scenario, the morphology of neuropodia could be summarized into three different states: ( Abdominal neuropodia of Amythas are similar to pinnuli, but this similarity is not constant since in large specimens uncini are not situated on the margin of neuropodia and remains in poorly developed furrows. The position of this tori/pinnuli change is also not stable (compare Fig. 5D and 5E). Therefore, it could be argued that neuropodia in Amythas represent a transitional stage between the plesiomorphic and apomorphic states: they are in the intermediate state, but this change is not so sharp and stable as in more apomorphic genera such as Ampharete, Amphicteis, Amage, Anobothrus, Grubianella, Jugamphicteis, Lysippe. Uncini also have a similar shape throughout.
This condition is at the base of the choice of Imajima et al. (2012) of proposing the new term "intermediate uncinigers". Imajima et al. (2012) also proposed that only those segments with neuropodial pinnules could be considered as "abdominal uncinigers" (Imajima et al. 2012: 76-77). Thus, in genera as Ampharete, Anobothrus, Grubianella, Lysippe, the body should be subdivided into thorax, transitional segments, and abdomen. This division, however, is unjustified for several reasons. For example, in Melinna Malmgren, 1866 and Orochi Reuscher, Fiege & Imajima, 2015 some posterior thoracic neuropodia are pinnuli. Therefore, by following their proposal, these posterior thoracic segments should be called transitional as well, despite they remain in the opposite direction.
Moreover, in some genera of Ampharetidae, such as Phyllocomus, Samythella and Amythas, the change of neuropodial shape is gradual, not sharp, and all segments (even thoracic ones) bear neuropodial pinnules (apomorphic state). By applying Imajima et al. (2012) scheme the logical base for subdivision the body on thorax and abdomen would be lost and the whole thorax area would remain undefined. Similar situation occurs in Terebellini and Artacamini (sensu Jirkov 2001) where some anterior abdominal segments can have neuropodia similar to posterior thoracic ones (double-row tori), while the rest abdominal segments have pinnuli like posterior segments of Ampharetidae or tori, but singlerow. The number of such anterior abdominal segments can vary from zero or few to several tens and sometimes almost all, but few posterior most abdominal segments have double-rows neuropodia. The same occurs in some Polycirrinae Malmgren, 1866 that have no parapodia at all but that should be called transitional as well, according to the Imajima et al. (2012) scheme. Given the above points, the Imajima et al. (2012) proposal of a new terminology seems unjustified and impossible to be applied in the practice. Therefore, it is here proposed to continue to use a simple and practical division, where the thorax area is defined by segments having notopodia with notochaetae and the abdomen by segments without notopodia with chaetae.
Other differences between Amythas and other genera lie in the shape of the prostomium. Amythas has a prostomium subdivided by a U-shaped groove and a rounded middle lobe as most of 'Ampharete-like' Ampharetinae. It differs from 'Amage-like' genera (such as Grubianella, Hypania and Samythopsis), which have an U-shaped groove, but also a middle lobe incised or even bearing horns.
Most of the genera with an Ampharete-like prostomium have peculiar traits: Eclysippe Eliason, 1955 and Auchenoplax Ehlers, 1887 have thorax sharply subdivided into two regions; Anobothrus and Sosane have modified notopodia in posterior thorax, Melinnampharete has transversal dorsal ridge between TC3 and TC4, Lysippe has lower lip enlarged, longitudinally grooved, Gnathampharete has numerous jaws. Only Amythas has none of these specific traits.
The above list of Amythas characters, when compared to those of similar genera should avoid further uncertainties in the identification of this genus, although many of these traits are not easily weighted by non-experts.