Redescription of Harmothoe spinosa Kinberg, 1856 (Polychaeta: Polynoidae) and related species from Subantarctic and Antarctic waters, with the erection of a new genus

During the expeditions ANT XIII/3 (1996) and ANT XV/3 (1998) to the Weddell Sea, a number of polynoid specimens resembling Harmothoe spinosa Kinberg, 1856 were collected. Due to contradictory information in the literature regarding the differentiating characters of H. spinosa and other related species, the type material of the former and of 27 nominal, similar species had to be examined in order to allow the correct identification of the specimens from the Weddell Sea. As a result of this study the validity of H. spinosa is confirmed, but this species has not been recorded in the Antarctic or Subantarctic region. Of the 27 similar species only nine are considered valid and redescribed herein, i.e. H. exanthema (Grube, 1858), H. fuligineum* (Baird, 1865), H. fullo* (Grube, 1878), H. antarctica* (McIntosh, 1885), H. magellanica* (McIntosh, 1885), H. crosetensis (McIntosh, 1885), H. acuminata* Willey, 1902, Antarctinoe gen. nov. ferox* (Baird, 1865), and A. spicoides* (Hartmann‐Schröder, 1986), with Antarctinoe being a new genus established herein. Species marked by an asterisk were also present in the collection of polynoids from the Weddell Sea. All records and geographical ranges given in the literature for H. spinosa and the other species not confirmed herein remain doubtful.


Introduction
During the cruises ANT XIII/3 (EASIZ I, 1996) and ANT XV/3 (EASIZ II, 1998) with R/ V Polarstern to the Weddell Sea Gutt 1997, 1999), a number of polynoids resembling Harmothoe spinosa Kinberg, 1856 were collected among other polychaetes. Although Hartman (1964) stated that H. spinosa is the first and best known polychaete, present in all sectors of Antarctica, the identification of the above-mentioned polynoids proved to be difficult. This is due to the fact that characters of H. spinosa are insufficently  segments 2, 4, 5, 7, on alternate segments to 23, 26, 29, and 32; first pair rounded, the following kidney-shaped, becoming oval more posteriorly; elytral margin with or without papillae; elytral surface with or without micro-and macrotubercles; anterior, inner part of surface (where elytra overlap) without tubercles; number of papillae and tubercles usually diminishing from anterior to posterior part of body. Dorsal cirri with cylindrical cirrophore and long style on segments lacking elytra; nodular dorsal tubercle on cirrigerous segments. Ventral cirri short (except on second segment), consisting of cirrophore and style on all segments. Prostomium bilobed, with distinct cephalic peaks and three antennae; ceratophore of median antenna in anterior notch, lateral antennae with ceratophores inserted ventrally to median antenna; a pair of palps present; two pairs of eyes, anterior pair situated anteroventrally beneath cephalic peaks or dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; pharynx with nine pairs of border papillae and two pairs of hooked jaws. Facial tubercle present.
First or tentacular segment with a pair of tentaculophores inserted laterally to prostomium, each bearing a dorsal and a ventral tentacular cirrus and one or two chaetae. Second or buccal segment with first pair of elytra, biramous parapodia, long ventral cirri, and a nuchal lobe.
Parapodia biramous, both rami with elongate acicular lobe; notopodia shorter and on anterodorsal side of larger neuropodia; neuropodia with digitiform supra-acicular process; tips of noto-and neuroacicula penetrating epidermis. Bases of neuropodia ventrally without tubercles between ventral cirrus and body wall. Notochaetae as stout as or stouter than neurochaetae, short and long notochaetae with distinct rows of spines and blunt tip; neurochaetae usually more numerous, with distinct rows of spines only distally; neurochaetae either all with bidentate tip with secondary tooth subdistally (tooth may be abraded, especially in older, larger animals) or some neurochaetae with bidentate tip and others with unidentate tip.
Nephridial papillae usually distinct from segment 5 to the end of the body. Pygidium with one pair of anal cirri similar in shape to dorsal cirri.

Remarks
Some species assigned herein to the genus Harmothoe Kinberg, 1856 have often been considered to belong to Hermadion Kinberg, 1856 in the past [see e.g. Hermadion fuligineum Baird, 1865, now considered as Harmothoe fuligineum, and Hermadion molluscum Ehlers, 1897and Hermadion ambiguum Ehlers, 1901, both synonyms of Harmothoe magellanica (McIntosh, 1885]. We checked the type species of Hermadion, i.e. the syntypes of Hermadion magalhaensi Kinberg, 1856 [SMNH-532 (one specimen) and SMNH-401 (three specimens)], and we can confirm that Harmothoe and Hermadion are two different genera, the main differentiating characters of Hermadion being the absence of cephalic peaks on the prostomium and neuropodia without a distal supra-acicular process.

Diagnosis
Elytral margin with scattered, digitiform papillae; elytral surface with conical microtubercles with blunt tip; near posterior margin few elongate, conical macrotubercles with blunt tip.

Description (based on lectotype)
Body with 38 segments. At anterior end ( Figure 1A), prostomium bilobed, with cephalic peaks; ceratophore of median antenna in anterior notch, lateral antennae inserted ventrally, styles of antennae missing in lectotype; anterior pair of eyes situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; palps missing.
Tentaculophores inserted laterally to prostomium, each with a single notochaeta and a dorsal and ventral tentacular cirrus with styles of cirri missing. Second segment with first pair of elytra (missing in lectotype), biramous parapodia, and long buccal cirri.

Remarks
Harmothoe spinosa with its elongate, conical macrotubercles is rather easily distinguishable from the other Harmothoe species described herein. The fact that the original description and figures by Kinberg (1856) were insufficient and that earlier authors (e.g. Ehlers 1897Ehlers , 1901aEhlers , 1901bEhlers , 1901cEhlers , 1913Bergströ m 1916) did not agree on the synonymy of this and related species created much confusion in the literature. Differences of H. spinosa and other morphologically similar species treated in this study are given in the ''Remarks'' section related to these species.

Distribution
So far only confirmed from the type locality, i.e. Magellan Strait (Magellan region, which is considered to be a cold-temperate region by Briggs 1996); not found among the material from Antarctic and Subantarctic origin examined herein.

Diagnosis
Elytral margin with scattered, digitiform papillae; elytral surface with conical microtubercles; near outer lateral and posterior margin pyriform to globose macrotubercles with or without distal papilla.
Tentaculophores inserted laterally to prostomium, each with a single notochaeta and a dorsal and ventral tentacular cirrus with styles of cirri papillate, abruptly tapering subdistally. Second segment with first pair of elytra, biramous parapodia, and long buccal cirri.

Diagnosis
Outer elytral margin densely set with filiform papillae, posterior margin without, or with few digitiform papillae; elytral surface with conical microtubercles, posterior half of elytra with numerous, globular to conical macrotubercles.

Description
The description is based on an additional specimen from the Weddell Sea (SMF 15162), since the lectotype of Hermadion fuligineum is in bad shape, although the diagnostic elytral characters are still visible. Body with 37 segments. At anterior end ( Figure 3A), prostomium bilobed, with cephalic peaks; ceratophore of median antenna in anterior notch, style of median antenna papillate, slightly inflated subdistally, then abruptly tapering; lateral antennae inserted ventrally with styles papillate, tapering; anterior pair of eyes situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; palps papillate, tapering.
Tentaculophores inserted laterally to prostomium, each with two notochaetae and a dorsal and ventral tentacular cirrus with styles of cirri missing. Second segment with first pair of elytra, biramous parapodia, and long buccal cirri.

Monro
Our comparison of the types of both species showed that Baird's species in fact belongs to Harmothoe (for differentiating characters between Hermadion and Harmothoe, see ''Remarks'' section related to the genus Harmothoe above), but that Harmothoe fuligineum is distinct from H. spinosa. Harmothoe fuligineum is characterized by a very dense fringe of papillae at the outer lateral elytral margin and by numerous globular to conical macrotubercles scattered in the posterior half of the elytron. In contrast, H. spinosa has only few papillae at the outer elytral margin and few elongate, conical macrotubercles occur near the posterior margin.
Among the additional material examined light and dark specimens could be distinguished, a phenomenon which had first been noted by Stiller (1996). Gambi et al. (2001) identified these specimens preliminarily as H. spinosa. According to the latter authors, the dark animals, while still alive, were characterized by a brownish to almost back body and iridescent elytra. Among the light animals, one specimen was found to brood eggs under its elytra, but dissection of the gonads of several specimens gave no clear evidence whether the dark or light morph is due to sexual dimorphism, since no difference could be noted in size and shape of the sexual products.

Diagnosis
Elytral margin with long and short digitiform papillae; elytral surface with conical microtubercles; few conical to cylindrical macrotubercles scattered near posterior margin; inner posterior margin with mound with very long, filiform papillae and some microtubercles, occasionally with few macrotubercles (mound less distinct in posterior elytra, but very long papillae still present).
Tentaculophores inserted laterally to prostomium, each with a single notochaeta and a dorsal and ventral tentacular cirrus with styles of cirri papillate, slightly inflated subdistally, then abruptly tapering. Second segment with first pair of elytra, biramous parapodia, and long buccal cirri.
Fifteen pairs of elytra, covering dorsum, on segments 2, 4, 5, 7, then on every second segment to 23, 26, 29, 32; last six segments cirrigerous; outer lateral and posterior margin of elytra with long and short digitiform papillae; elytral surface covered by conical microtubercles, few conical to cylindrical macrotubercles scattered near posterior margin; inner posterior margin with a mound with very long, filiform papillae and some microtubercles, occasionally also with few macrotubercles (mound becoming less distinct from anterior to posterior elytra, but group of long, filiform papillae still present on posterior elytra) ( Figure 4B-D). Cirrigerous segments with distinct dorsal tubercles; dorsal cirri with cylindrical cirrophore, style papillate, abruptly tapering subdistally. Ventral cirri with smooth, tapering style.

Remarks
We re-examined the syntypes of Polynoe fullo Grube, 1878 which were also checked by Augener (1932a (1897) and Augener (1932a) regarded Polynoe fullo as a synonym of H. spinosa. But our investigations of the type material revealed that both species are clearly different with Harmothoe fullo having elytra with few conical to cylindrical macrotubercles scattered near the posterior margin and with a very characteristic mound at the inner posterior margin with very long, filiform papillae and some microtubercles, occasionally with few macrotubercles.
Harmothoe spinosa in contrast never shows such a mound and is characterized by few elongate, conical macrotubercles with blunt tip near the posterior margin.
Although the type material of Harmothoe gourdoni Gravier, 1911 was not available in the MNHN in Paris and is probably lost (cf. F. Pleijel, personal communication), the description and figures given by Gravier (1911aGravier ( , 1911b are clear and leave no doubt that this species is a junior synonym of Harmothoe fullo (Grube, 1878) . Bergströ m (1916) suggested that Harmothoe gourdoni might be synonymous with Harmothoe spinosa, which at that time consisted already of a mixture of different species, among them Polynoe fullo.
Harmothoe fullo might be confused with Harmothoe antarctica (McIntosh, 1885), both species sharing elytra with a characteristic mound at the inner posterior margin. But the spherical macrotubercles present at the centre of the elytron and near the posterior margin together with the always short papillae on the mound are characteristic for H. antarctica, while H. fullo shows very long, filiform papillae on the mound and less numerous, conical to cylindrical macrotubercles which occur only near the posterior margin.

Diagnosis
Elytral margin with digitiform papillae; elytral surface with conical microtubercles; patch of smaller, spherical macrotubercles present centrally and numerous, larger, spherical macrotubercles near posterior margin; inner posterior margin with a mound with microand macrotubercles and short papillae (mound becoming less distinct in posterior elytra).

Description (based on lectotype of Lagisca antarctica)
Body with 38 segments. At anterior end ( Figure 5A), prostomium bilobed, with cephalic peaks; ceratophore of median antenna in anterior notch, style of median antenna missing; lateral antennae inserted ventrally with styles papillate, tapering; anterior pair of eyes situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; palps papillate, tapering (right palp missing).
Tentaculophores inserted laterally to prostomium, each with a single notochaeta and a dorsal and ventral tentacular cirrus with styles of cirri papillate, abruptly tapering subdistally. Second segment with first pair of elytra, biramous parapodia, and long buccal cirri.

Remarks
Since Harmothoe antarctica (McIntosh, 1885), originally assigned to Lagisca, was synonymized with H. spinosa by Ehlers (1913), this species is probably also hidden in the subsequent descriptions of H. spinosa of many authors. But, in contrast to other species confused with H. spinosa, the characteristic features of H. antarctica

Diagnosis
Outer lateral elytral margin and posterior half of surface with few digitiform papillae; anterior half of elytra covered with conical microtubercles, number of microtubercles and surface covered variable (papillae and microtubercles missing in posterior elytra).

Description (based on lectotype of Lagisca magellanica)
Body with 42 segments. At anterior end ( Figure 6A), prostomium bilobed, with cephalic peaks; ceratophore of median antenna in anterior notch, lateral antennae inserted ventrally, styles of antennae missing in lectotype; anterior pair of eyes situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; palps missing. [Prostomial and tentacular appendages present in some of the paralectotypes: styles of antennae and tentacular cirri papillate, abruptly tapering subdistally; palps papillate, tapering.] Tentaculophores inserted laterally to prostomium, each with a single notochaeta and a dorsal and ventral tentacular cirrus with styles of cirri missing. Second segment with first pair of elytra (missing in lectotype), biramous parapodia, and long buccal cirri.
Nephridial papillae distinct from segment 12 (in other investigated specimens distinct generally from segment 6).

Remarks
As described above for the lectotype of L. magellanica, the area covered by the microtubercles on the elytral surface and by the papillae at the elytral margin varies depending on the position of the elytron on the body. Furthermore, we observed also an individual variability regarding these characters in the investigated type and additional material, some specimens showing a larger area covered by the tubercles in anterior elytra, while in others the area covered is reduced to an anterior patch.
Ehlers (1897) considered L. magellanica and the two subspecies murrayi and grubei to be junior synonyms of H. spinosa Kinberg, 1856. But, as we demonstrated above, H. spinosa mainly differs from H. magellanica in the presence of elongate, conical macrotubercles near the posterior elytral margin and microtubercles covering more or less the complete elytral surface, while in H. magellanica macrotubercles are missing, and microtubercles are present only in the anterior half of the elytron.

Distribution
Magellan region: Magellan Strait and South Patagonia; Subantarctic region: Kerguelen Islands (type locality) and Prince Edward Island; Antarctic region: South Georgia, Antarctic Peninsula, and Weddell Sea (cf. also Stiller 1996).

Diagnosis
Elytral margin with digitiform papillae getting shorter towards posterior margin; elytral surface densely covered by microtubercles, being conical, elongate with blunt tip in anterior half, becoming gradually larger towards posterior margin, with blunt to slightly multifid tip.

Description (based on lectotype of Lagisca crosetensis)
Body with 36 segments. At anterior end ( Figure 8A), prostomium bilobed, with cephalic peaks; ceratophore of median antenna in anterior notch, style of median antenna missing; lateral antennae inserted ventrally with styles missing; anterior pair of eyes situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; palps missing.
Tentaculophores inserted laterally to prostomium, each with one or two notochaetae and a dorsal and ventral tentacular cirrus with styles of cirri papillate, abruptly tapering subdistally (style of left and right dorsal and left ventral tentacular cirri missing). Second segment with first pair of elytra, biramous parapodia, and long buccal cirri.
Fifteen pairs of elytra, covering dorsum, on segments 2, 4, 5, 7, then on every second segment to 23, 26, 29, 32; last four segments cirrigerous; outer lateral margin of elytra with digitiform papillae getting shorter towards posterior margin; posterior half of elytral surface also covered by papillae; elytral surface densely covered by microtubercles, being conical, elongate with blunt tip in anterior half, becoming gradually larger towards posterior margin, with blunt to slightly multifid tip ( Figure 8B-D). Cirrigerous segments with distinct dorsal tubercles; dorsal cirri with cylindrical cirrophore, style missing on all segments. Ventral cirri with smooth to papillate, tapering style.

Diagnosis
Elytral margin with digitiform papillae; near anterior elytral margin surface with conical, pointed, or blunt microtubercles, central and posterior part of elytron with numerous, pointed, thorn-to spine-shaped macrotubercles getting larger towards posterior margin.

Description (based on lectotype of Harmothoe crosetensis acuminata)
Body with 34 segments. At anterior end ( Figure 9A), prostomium bilobed, with cephalic peaks; ceratophore of median antenna in anterior notch, style of median antenna missing; lateral antennae inserted ventrally with papillate, tapering styles; anterior pair of eyes situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; palps papillate, tapering.
Tentaculophores inserted laterally to prostomium, each with a single notochaeta and a dorsal and ventral tentacular cirrus with styles of cirri regenerating or broken (style of right dorsal tentacular cirrus missing). Second segment with first pair of elytra, biramous parapodia, and long buccal cirri.

Remarks
Since the differences between H. crosetensis and H. acuminata have been extensively discussed above, we focus here only on the characters separating H. acuminata from H. serrata Day, 1963 (see also above). Both species share papillate elytral margins and surface and pointed, thorn-or spine-shaped elytral macrotubercles. But in H. serrata thorn-shaped macrotubercles are only present near the posterior elytral margin, while in H. acuminata macrotubercles are thorn-to spine-shaped, covering the central and posterior part of the elytron and becoming larger towards the posterior margin. Among the Antarctic species described herein, the two species belonging to the new genus Antarctinoe (see below), i.e. A. ferox (Baird, 1865) and A. spicoides (Hartmann-Schröder, 1986), are also characterized by thorn-shaped, pointed macrotubercles. However, in these two species the characteristic papillae occurring at the elytral margin and the surface of both H. acuminata and H. serrata are missing. Furthermore, species belonging to Antarctinoe gen. nov. differ from Harmothoe in the characteristic length and orientation of the notochaetae (for more details, see below) and the exclusive presence of unidentate neurochaetae.

Distribution
Antarctic region: South Georgia and Graham region (Bergströ m 1916), Weddell Sea, and Cape Adare, Ross Sea (type locality).

Diagnosis
Body dorsoventrally flattened, short, with fewer than 50 segments; dorsum covered only partly by elytra, except first two to three pairs. Fifteen pairs of elytra on segments 2, 4, 5, 7, on alternate segments to 23, 26, 29, and 32. Dorsal cirri with cylindrical cirrophore and long style on segments lacking elytra; pair of nodular dorsal tubercles on cirrigerous segments. Ventral cirri short (except on second segment), consisting of cirrophore and style on all segments.
Prostomium bilobed, with distinct cephalic peaks and three antennae; ceratophore of median antenna in anterior notch, lateral antennae with ceratophores inserted ventrally to median antenna; a pair of palps present; two pairs of eyes, anterior pair situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; pharnyx with nine pairs of border papillae and two pairs of hooked jaws. Facial tubercle present.
First or tentacular segment with a pair of tentaculophores inserted laterally to prostomium, each bearing a dorsal and a ventral tentacular cirrus and a single notochaeta. Second or buccal segment with first pair of elytra, biramous parapodia, long ventral cirri, and a nuchal lobe.
Parapodia biramous, both rami with elongate acicular lobe; noto-and neuropodia of about same size; neuropodia with digitiform supra-acicular process; tips of noto-and neuroacicula penetrating epidermis. Bases of neuropodia ventrally without tubercles between ventral cirrus and body wall. Notochaetae of several kinds (depending on the species); all known species with very long, stout notochaetae, some projecting laterally, the majority projecting dorsally to meet mid-dorsally; longest stout notochaetae distinctly longer than longest neurochaetae (up to about twice the length).
Neurochaetae falcate with unidentate tip. Nephridial papillae distinct from segment 6 to the end of the body. Pygidium with one pair of anal cirri.

Remarks
In our material from the Weddell Sea, we found specimens belonging to two species having very long, stout notochaetae projecting dorsally to meet mid-dorsally, unidentate neurochaetae, and the same ornamentation of the elytra, but differing in the number of notochaetal types and the shape of the notochaetal tips. In the literature specimens showing only short and very long stout notochaetae with more or less blunt tips, were referred to Hermadion or Harmothoe ferox, to Eunoe spica, or to Harmothoe crosetensis among others. Specimens with up to four kinds of notochaetae (i.e. short notochaetae with pointed tip; very long stout notochaetae with pin-like tip; fine, slender notochaetae with pointed to pinlike tip; and capillary notochaetae with pointed tip; with the presence of the latter two types occurring only in large specimens) were described as Harmothoe (Eunoe) spica spicoides or H. crosetensis (for extensive synonymy see below).
We establish herein the new genus Antarctinoe gen. nov. for the two species A. ferox (Baird, 1865) and A. spicoides (Hartmann-Schrö der, 1986), described below, and characterized mainly by the presence of 15 pairs of elytra, fewer than 50 segments, prostomium with cephalic peaks, neuropodia with a supra-acicular process, very long stout notochaetae mostly orientated dorsally to meet mid-dorsally, and unidentate neurochaetae.
Antarctinoe gen. nov. is easily separated from Harmothoe Kinberg, 1856 due to the exclusive presence of unidentate neurochaetae, never being distinctly bidentate as in Harmothoe. Furthermore, the very long stout notochaetae orientated to protect the dorsum are a unique character of Antarctinoe enabling the further distinction between this genus and Harmothoe, and also from Eunoe Malmgren, 1865. To our knowledge, in the latter two genera the longest stout notochaetae are never distinctly longer than the longest neurochaetae (usually shorter or of about same length) and mostly orientated laterally (see Barnich and Fiege 2000, 2003, 2004. The differentiation between Antarctinoe and Hermadion Kinberg, 1856 is rather easily demonstrated. Our investigations of a syntype of Hermadion magalhaensi Kinberg, 1856 (SMNH-532), i.e. the type species of the genus, showed that Hermadion is characterized by rounded anterior prostomial edges without cephalic peaks and acicular neuropodial lobes without supra-acicular process, while Antarctinoe shows a prostomium with cephalic peaks and neuropodia with a distinct supra-acicular process distally at the acicular lobe. Furthermore, the characteristic size and orientation of the notochaetae as described above for Antarctinoe are not found in Hermadion.

Etymology
The name refers to the geographical area (Antarcti-) and to the closely related genus Eunoe (-noe); gender feminine.

Diagnosis
Elytral margin smooth; elytral surface with conical microtubercles with blunt to pointed tip; thorn-shaped macrotubercles scattered in posterior half of elytron. Long, stout notochaetae projecting dorsally to meet mid-dorsally with rows of spines distally and pin-like tip; large specimens with additional fine, slender notochaetae with pointed to pin-like tip, and capillary notochaetae with pointed tip. Neurochaetae falcate, unidentate.

Description
The description is based on the holotype of Harmothoe (Eunoe) spica spicoides, but notochaetal characters are complemented by an additional specimen from the Weddell Sea (SMF 15165); see ''Remarks'' section below.
Body with 42 segments. At anterior end ( Figure 10A), prostomium bilobed, with cephalic peaks; ceratophore of median antenna in anterior notch, style missing; lateral antennae inserted ventrally, styles missing; anterior pair of eyes situated dorsolaterally on widest part of prostomium, posterior pair dorsally near hind margin of prostomium; palps missing.
Tentaculophores inserted laterally to prostomium, each with a single notochaeta and a dorsal and ventral tentacular cirrus with styles papillate, slightly inflated subdistally, then abruptly tapering (styles missing, except for left dorsal tentacular one). Second segment with first pair of elytra, biramous parapodia, and long buccal cirri.
Notochaetae of several kinds depending on the size of the specimen; small to mediumsized specimens with short, stout notochaetae with rows of spines and more or less pointed tip and long, stout notochaetae projecting dorsally to meet mid-dorsally with rows of spines distally and pin-like tip ( Figures 10E, F, 11B, C); large specimens additionally with fine, slender notochaetae with rows of spines distally and pointed to pin-like tip, and capillary notochaetae with rows of spines distally and pointed tip ( Figure 11A-E). Neurochaetae falcate with unidentate tip ( Figure 10G).
Notochaetae of two kinds independent of size of individual: short, stout notochaetae with rows of spines and more or less pointed tip, and long, stout notochaetae projecting dorsally to meet mid-dorsally with blunt tip and shaft usually smooth (in smaller specimens few rows of spines present distally) ( Figure 12F-H).
Neurochaetae falcate with unidentate tip; one or two neurochaetae with subdistal slit suggesting presence of secondary tooth (but never distinctly bidentate as in Harmothoe) ( Figure 12I-K).

Remarks
Although the type specimen of Hermadion ferox Baird, 1865 is in bad shape with no remaining elytra, the notochaetal characters together with the description given in Baird (1865) leave no doubt regarding its authenticity. Unfortunately Baird does not clearly state if his description is based on a single specimen, i.e the holotype, or on syntypes.
The type material of Hermadion rouchi Gravier, 1911 is not available in the MNHN in Paris and thus probably lost (cf. F. Pleijel, personal communication). But the description and figures by Gravier (1911b) are sufficient to synonymize this species with Antarctinoe ferox. This was also recognized by Monro (1929) who was the first to synonymize Hermadion rouchi with Hermadion ferox, and by Uschakov (1962) who considered the species to belong to the genus Harmothoe. According to Monro (1930), Hermadion ferox is difficult to differentiate from Harmothoe crosetensis sensu Bergströ m (1916) which is considered herein to be synonymous with Harmothoe acuminata Willey, 1902 (see above). But Willey's description and figures of his subspecies Harmothoe crosetensis acuminata are rather confusing. Our investigations of the respective type material showed that Harmothoe acuminata is easily distinguished from Antarctinoe ferox, the former having some middle neurochaetae with distinctly bidentate tip, short and long notochaetae orientated mostly laterally, and elytra with papillate margin and numerous elongate, pointed, spine-shaped macrotubercles, while A. ferox differs in its unidentate neurochaetae, very long and stout notochaetae orientated mostly mid-dorsally, and elytra with smooth margin and smaller, curved, thorn-shaped macrotubercles.
Besides the similar elytral ornamentation, Antarctinoe spicoides and A. ferox share the same type of short notochaetae and the same characteristic orientation of the very long, stout notochaetae pointing mostly mid-dorsally. But A. spicoides can be clearly separated from A. ferox by the presence of pin-like tips in the long, stout notochaetae and larger specimens showing additional slender and capillary notochaetae, while A. ferox is characterized by only a single kind of long, stout notochaetae with blunt tip and usually smooth shaft.
Within the specimens from station ANT XIII/3, 39/012 we observed two different morphs independent of the size of the individuals. Some specimens show pigmented transverse bands on the dorsum and at the elytral margin (in ethanol) together with very long, stout notochaetae extending to the elytrophores or dorsal tubercles of the opposite side (the holotype and the figured specimen belonging also to this morph). Other specimens from the same station are more or less unpigmented (in ethanol) and have long stout notochaetae extending only to about the mid-dorsum. Since the specimens were not sexually reproductive (i.e. dissection of gonads showed no clear distinction between immature sperm or eggs) we do not know if the occurrence of the two morphs is due to sexual dimorphism.
Identification key to the species covered herein