On the genus Halirages ( Crustacea , Amphipoda ) , with the description of two new species from Scandinavia and Arctic Europe

A new common deep-sea species of Halirages Boeck, 1871 closely related to H. qvadridentatus G.O. Sars, 1877, H. cainae sp. nov., is described after specimens collected in the Norwegian Sea during the MAREANO 2009-111 cruise. Examination of the syntypes of H. elegans Norman, 1882 demonstrates that Norman's species is a junior synonym of H. qvadridentatus G.O. Sars, 1877 and that the species usually named H. elegans in literature was actually undescribed. The name H. stappersi sp. nov. is proposed for that species. A key to and a checklist of Halirages species is given.


Introduction
According to Barnard & Karaman (1991), the amphipod genus Halirages Boeck, 1871 includes 7 shallow-and deep-water species, all from Arctic and sub-Arctic Seas: H. caecus Kamenskaya, 1980, H. elegans Norman, 1882, H. fulvocinctus (M. Sars, 1859), H. gorbunovi Gurjanova, 1946, H. mixtus Stephensen, 1931, H. nilssoni Ohlin, 1895and H. qvadridentatus G.O. Sars, 1877, but in many respects this genus remains imperfectly studied.During the second leg of the MAREANO 2009-111 cruise (29 Sep. 2009-11 Oct. 2009) on the eastern slope of the Norwegian deep Sea, a large crimson Halirages species was observed with the video platform Campod (see e.g.Anonym 2006;Buhl-Mortensen et al. 2009) and collected in several trawl and sledge hauls.That common species proved to be closely related to, albeit distinct from, the well-known Scandinavian bathyal species Halirages qvadridentatus.At fi rst it was assumed that the taxon at hand could be H. elegans, a species inadequately described after specimens from bathyal Faeroe waters (Norman 1882).However, examination of the type specimens of H. elegans revealed that they consisted of characteristic H. qvadridentatus, as already assumed by Della Valle (1893) and Stebbing (1906).The crimson Halirages also exhibited some similarities with two further deep-sea species from the Arctic Ocean: H. gorbunovi from the eastern part of the Nansen Basin and H. caecus from the Canadian Basin.The published accounts of these two species from distant seas are poor and based on badly mutilated specimens, but did not fi t with the crimson Halirages from the Norwegian Sea.Therefore the latter is described herein as a new species, H. cainae sp.nov.A further consequence of the synonymization of H. elegans with H. qvadridentatus is that the North-European species reported as H. elegans by Stappers (1911) and Stephensen (1931) actually belongs to another species, which has no valid name.That species is named herein H. stappersi sp.nov.Finally, a key to and a checklist of Halirages species is given.

Material and Methods
The specimens and appendages were examined with a Wild 181300 dissecting microscope, and a DML Leica compound microscope, both equipped with a drawing tube.Pencil drawings were scanned and afterwards inked with the software Adobe ® Illustrator ® 11.0 on an A3 drawing table (Wacom ® Intuos3 12x19), using the method described by Coleman (2003Coleman ( , 2009)).A complete description is given for H. cainae sp.nov., which was a previously unknown species.A description of the parts of the syntypes of H. elegans visible without further dissection is provided to demonstrate its identity with H. qvadridentatus.A description of H. stappersi sp.nov.expanding the account of Stappers (1911 under the name of H. elegans) with further information is given, but no further dissection of the unique type specimen was carried out, because this was considered unnecessary and destructive.A reproduction of all the fi gures of Stappers (1911) including those of lost appendages is also given.These fi gures are no longer under copyright.Jean Hubert Louis Stappers was Belgian and his book was published by the 'Imprimerie Scientifi que Charles Bulens' in Belgium.The Belgian law on copyright and related rights of 30 Jun. 1994, art. 2 (http://www.wipo.int/wipolex/en/text.jsp?fi le_id=125150 [accessed 24 Nov. 2011]) states that the term of protection for copyrighted material expires 70 years after the death of the author.Jean Hubert Louis Stappers died on 30 Dec. 1916 (http://www.hasel.be/nl/subjects/976/stappers-louis-1883-1916. html [accessed 24 Nov. 2011]]), i.e. more than 70 years ago.The following abbreviations are used in the lists of material examined, keys and in the captions of the table and the fi gures: A1 = antenna 1; A2 = antenna 2; Ep1-Ep3 = epimeral plates 1-3; Gn1-Gn2 = gnathopods 1-2; Md = mandible; Mx1 = maxilla 1; Mx2 = maxilla 2; Mxp = maxilliped; P3-P7 = pereiopods 3-7; U1-U3 = uropods 1-3.A numerical name is given for the articles of the palp of mouthparts: articles 1-3 for mandibular palp, articles 1-2 for palp of maxilla 1, articles 1-4 for palp of maxilliped.In the description, the term 'tooth' is used for non-articulated, pointed ectodermic structures, the term 'spine' for stout, articulated infl exible structures, and the term 'seta' for slender, articulated fl exible structures (see d'Udekem d'Acoz 2010).The setae of the mandibular palp are named according to the nomenclature proposed by Lowry & Stoddart (1993).Concerning measurements, the body length is given from tip of rostrum to tip of telson.When station depths are originally given in fathoms, the original unit is retained but a conversion in metres is given between brackets, in assuming that the fathoms were international fathoms (1.8288 metres).However different kinds of fathoms were in use in the nineteenth century, so that the real depth in metres may possibly be slightly different.The following abbreviations and acronyms are used for research programs and scientifi c institutions: MAREANO = Marine AREAdatabase for NOrwegian coast and sea areas; NHM: the Natural History Museum, London, UK (previously British Museum, Natural History); RBINS: Royal Belgian Institute of Natural Sciences, Brussels, Belgium; ZMBN, Zoologisk Museum, Universitetet i Bergen, Naturhistorie, Bergen, Norway.

Gender
Halirages is considered as masculine in older literature, but often as feminine in recent faunistic papers.
There is no apparent reason for this change and the issue needs clarifi cation.The second part of the name is a (very liberal) derivation from the Greek (see section etymology), so this case should be decided under Article 30.1 of ICZN (1999), which concerns the gender of names formed from Latin or Greek words.Article 30.1.4.2.states that a genus-group name that is or ends in a word of common or variable gender (as it is the case of words ending in -es) is to be treated as masculine unless its author, when establishing the name, stated that it is feminine or treated it as feminine in combination with an adjectival speciesgroup name.Since Boeck (1871) gave masculine adjectival names to H. bispinosus (Spence Bate, 1857), H. tridentatus Bruzelius, 1859 and H. fulvocinctus (M.Sars, 1859) (and an adjectival name, which can be either masculine or feminine to H. borealis Boeck, 1871), Halirages should be treated as masculine.

Remarks
The species accepted in Halirages are the two new species described herein and those included by Barnard & Karaman (1991).The latter authors transferred Halirages bungei Gurjanova, 1951to Paracalliopella Tzvetkova & Kudrjaschov, 1975, H. megalops (Buchholz, 1874) to Apherusa Walker, 1891, and H. huxleyanus (Spence Bate, 1862), H. batei (Cunningham, 1871) and H. regis (Stebbing, 1914) to Austroregia Barnard, 1989. Stephensen (1931) pointed out that the borderline between the genera Halirages and Apherusa is fuzzy.His remark remains more pertinent than ever, and a more consistent delimitation and defi nition of Arctic and sub-Arctic calliopiid genera would be more than welcome.becoming strong and narrowly spaced near tip; article 3 falciform, with row of D3-setae on distal 0.8 of posterior border, with apical tuft of E3-setae.

Halirages cainae
MAXILLA 1. (Fig. 3G-L) Inner plate with 9 plumose setae, of which the length signifi cantly increases towards tip; outer plate with 10 spines on left side and 8 on right side (they include deeply bifurcate spines and spines with 1-3 strong denticles on posterior border); palp well developed, with broad article 2; left article 2 with row of long styliform marginal spines and margino-facial setae (most of them forming a row); right article 2 with row of stout conical marginal spines (partly fused with article 2), with 2 longer anterodistal spines (most upper one arising medially), with margino-facial row of setae.
MAXILLA 2. (Fig. 3M) Basal part with 2 short anterior setae (one is lost on illustrated Maxilla 2); plates rather narrow; inner plate with marginal and margino-facial row of setae respectively on distal 0.7 and 0.6 of posterior border, with microtrichs on proximal 0.3 of posterior border; outer plate with upper border straight, with one tiny anteroproximal medial seta, with microtrichs on upper subdistal border, with about 6 short anterodistal setae (2 are lost on illustrated Maxilla 2), with double row of strong distal setae.
GNATHOPOD 1. (Fig. 4B-C) Coxa with anteroventral corner forming a square angle (not produced into a tooth pointing forwards), with 10 weak crenulations along ventral margin; carpus 3.8 x as long as wide, as long as basis, anterior border not setose (except for distal tuft of seta); propodus 2.3 x as long as wide, 0.71 x as long as carpus; palm sharply denticulate, with marginolateral row of thin setae; palmar part of propodus 0.31 x as long as propodus; dactylus dentate all along its posterior border.
GNATHOPOD 2. (Fig. 4D) Coxa square, with 10 weak crenulations along ventral margin; carpus 5.0 x as long as wide, as long as basis, with anterior border scarcely setose to glabrous (except for distal tuft of seta); propodus 2.5 x as long as wide, 0.63 x as long as carpus; palm sharply denticulate, with marginolateral row of thin setae; palmar part of propodus 0.25 x as long as propodus; dactylus dentate all along its posterior border.
PEREIOPOD 3. (Fig. 5A) Coxa square with about 10 weak and in some cases indistinct crenulations; leg weakly spinose/setose; basis anteriorly distinctly concave and posteriorly distinctly convex, with setae on its two borders; carpus 8.9 x as long as wide, 1.8 x as long as merus; propodus 12.6 x as long as wide, 2.1 x as long as merus; dactylus 0.41 x as long as propodus, 0.88 x as long as merus.
PEREIOPOD 4. (Fig. 5B-C) Coxa broad and posteriorly produced into a bluntly triangular protrusion, with ventral margin with 10 weak crenulations; leg weakly spinose/setose, slightly longer than pereiopod 3; basis anteriorly concave and posteriorly convex, with setae on its two borders; carpus 9.0 x as long as wide, 1.8 x as long as merus; propodus 13.5 x as long as wide, 2.2 x as long as merus, 1.2 x as long as propodus of pereiopod 3; dactylus 0.39 x as long as propodus, 0.87 x as long as merus.
PEREIOPOD 5. (Fig. 6A-B) Pereiopod 5 < pereiopod 6 < pereiopod 7; posterior lobe of coxa distinctly longer than anterior lobe; leg weakly spinose/setose; basis elliptic, 1.3 x as long as wide, anterior border with 8 styliform spines and sparse thin setae, distally without tooth, posterior border with 9 very low crenulations, posterodistal border rounded and smooth, with 2 long styliform posterodistal medial spines; ischium without anterodistal tooth; carpus 11.6 x as long as wide, 1.7 x as long as merus; propodus 19 x as long as wide, 1.9 x as long as merus; dactylus 11 x as long as wide, 0.33 x as long as propodus, 0.63 x as long as merus.PEREIOPOD 6. (Fig. 6C-D) Posterior lobe of coxa considerably longer than anterior lobe; leg weakly spinose/setose; basis elliptic, 1.3 x as long as wide, anterior border with 9 styliform spines (one rubbed off on illustrated basis) and sparse thin setae, distally without tooth, posterior border with 14 very low crenulations, posterodistal border rounded and smooth, with 2 long styliform posterodistal medial spines; ischium without anterodistal tooth, with anterodistal spine; carpus 11.5 x as long as wide, 1.6 x as long as merus; propodus 19 x as long as wide, 1.7 x as long as merus; dactylus 13 x as long as wide, 0.29 x as long as propodus, 0.51 x as long as merus.
PEREIOPOD 7. (Fig. 6D-E) Coxa small and elliptic; large coxal gill present; leg weakly spinose/setose; basis with anterior and posterior border straight and converging towards tip, 1.2 x as long as wide, anterior border with 5 styliform spines and a few thin setae, distally without tooth, posterior border with 18 serrations, posterodistal border with 3 serrations; junction between posterior and posterodistal border bluntly angular; ischium without anterodistal tooth; carpus 11.2 x as long as wide, 1.6 x as long as merus; propodus 18.4 x as long as wide, 1.8 x as long as merus, 1.02 x as long as propodus of P6, 1.30 x as long as propodus of P5; dactylus 12 x as long as wide, 0.28 x as long as propodus, 0.49 x as long as merus.
UROPOD 3. (Fig. 7F) Peduncle with 4 distolateral dorsal spines, with dorsomedial and ventromedial border spinose; outer ramus 0.93 x as long as inner ramus, with about 47 lateral irregular-sized (most small and slender) spines (several lost on illustrated uropod), with at least 46 medial irregular-sized (most small and slender) spines; inner ramus 2.1 x as long as peduncle, without distinct medio-proximal bulging, with 49 lateral spines (most small and slender; some lost on illustrated uropod) and at least 15 plumose setae, with 49 medial slender irregular-sized spines (most small and slender; some lost on illustrated uropod) and at least 9 plumose setae.TELSON.(Fig. 4E) Triangular, with border convex, distally produced into a single tooth, without setules.
COLOUR PATTERN.(Fig. 1) Uniformly crimson, eyes dull reddish pink.In alcohol, the red pigment persists longer on the oral fi eld and the gnathopods.
MOUTHPARTS.Typical for the genus but not described for the poor quality of the dissected material.
GNATHOPOD 1. (Fig. 8B-D) Coxa with anteroventral corner produced into a small tooth pointing forwards, with 18-20 small serrations along ventral margin; carpus 4.4 x as long as wide, as long as basis, anterior border not setose; propodus 2.4 x as long as wide, 0.66 x as long as carpus; palmar part of propodus 0.23 x as long as propodus.
GNATHOPOD 2. (Fig. 8B, E) Coxa rectangular, with about 16 crenulations along ventral margin; carpus 4.8 x as long as wide, as long as basis, with anterior border not setose; propodus 2.5 x as long as wide, 0.59 x as long as carpus; palmar part of propodus 0.23 x as long as propodus.
PEREIOPOD 7. (Figs 8A-B, 9C-D) Coxa small and elliptic; leg weakly spinose/setose; basis with anterior and posterior border straight and converging towards tip, 1.3 x as long as wide, anterior border with 6 slender spines and no setae, distally without tooth, posterior border with 22-37 serrations, posterodistal border with 3-5 serrations; junction between posterior and posterodistal border sharply angular; ischium without anterodistal tooth; carpus 8.8 x as long as wide, 1.4 x as long as merus; propodus 13 x as long as wide, 1.6 x as long as merus.
TELSON.(Fig. 9F-G) Triangular, with border convex, distally produced into 3 teeth of which the median one is by far the longest, without setules.
BODY LENGTH.The largest syntype of H. elegans is 30 mm long.

Remarks
The original description of H. elegans, based on specimens from the Faeroe Islands, was meager and devoid of illustrations (Norman 1882), so that it was impossible to fi gure out exactly what the species looked like and how it differed from congeners.Stappers (1911), reluctantly followed by Stephensen (1931), applied the name H. elegans to a species very distinct from H. qvadridentatus and in many respects closer to H. fulvocinctus.Examination of the type material of H. elegans demonstrated that it consisted of specimens of H. qvadridentatus, confi rming the assumptions of Della Valle (1893) and Stebbing (1906).A description and fi gures of these type specimens are given herein to demonstrate their synonymy and a new name is proposed for the species described in Stappers (1911) and Stephensen (1931) in the next section.Stephensen 1938: 237, (key), 241. -Gurjanova 1946: 287 (discussion).-Yashnov 1948: 641, pl. 78 (after Stephensen 1931).-Gurjanova 1951: 605, 607-608, fi g. 410 (after Stappers 1911).

Etymology
Halirages stappersi sp.nov. is dedicated to the memory of Jean Hubert Louis Stappers (1883Stappers ( -1916)), who collected the holotype of the species during the Arctic Campaign of 1907 of the Duc d'Orléans on the RV Belgica (see Barr 2010) and described it accurately, but under the name of Halirages elegans Norman, 1882.The name is a genitive.

Description
HEAD. (Fig. 10A-B) Rostrum feeble; anterior lobe of head very bluntly subquadrate (almost rounded), posteriorly followed by narrow sinus; ventral lobe of head acute, pointing forward, not denticulate; eye large, subquadrate, with fully developed ommatidia, pigmentation retained after being stored in alcohol for a century.
LOWER LIP. (Fig. 10D) With narrow mandibular processes and broad outer lobes.
MAXILLA 1. (Fig. 10 F-G) Inner plate with 8 plumose setae, the length of which the size signifi cantly increases towards tip; outer plate with 9 denticulate spines; palp well developed, with broad article 2; left article 2 with row of long styliform marginal spines and row of margino-facial setae; right article 2 with distal margin dentate, with 2 freely articulated anterodistal spines, with margino-facial row of welldeveloped setae.
GNATHOPOD 1. (Figs 10A, 11A) Coxa with anteroventral corner produced into a tooth pointing forwards, with 12-15 strong serrations along ventral margin; carpus 3.9 x as long as wide, almost as long as basis, anterior border without setae (except for distal tuft of seta); propodus 2.6 x as long as wide, 0.78 x as long as carpus; palm denticulate, with row of thin setae; palmar part of propodus 0.25 x as long as propodus; dactylus dentate.
GNATHOPOD 2. (Figs 10A, 11B) Coxa broadly rectangular, with about 13 distinct serrations along ventral margin; carpus 3.8 x as long as wide, distinctly shorter than basis, with anterior border setose; propodus 2.6 x as long as wide, 0.74 x as long as carpus; palm denticulate; palmar part of propodus about 0.22 x as long as propodus; dactylus dentate.PEREIOPOD 3. (Fig. 11C) Coxa slightly longer than broad, with anterior and posterior border parallel, distally rounded, with ventral border serrate; leg distinctly spinose/setose; basis anteriorly weakly concave and posteriorly weakly convex; carpus 7.0 x as long as wide, 1.3 x as long as merus; propodus 8.0 x as long as wide, 1.5 x as long as merus; dactylus 0.39 x as long as propodus, 0.58 x as long as merus.
PEREIOPOD 7. (Fig. 11G-H) Leg distinctly spinose/setose; basis elliptic, 1.5 x as long as wide, anterior border setose, with small but sharp anterodistal tooth, posterior and posterodistal border with 29 serrations (distal serrations irregularly shaped); junction between posterior and posterodistal border bluntly angular; ischium with small but sharp anterodistal tooth; carpus 5.8 x as long as wide, 0.94 x as long as merus; propodus 13 x as long as wide, 1.2 x as long as merus; dactylus 7 x as long as wide, 0.36 x as long as propodus, 0.44 x as long as merus.
TELSON.(Fig. 12E-F) Triangular, distally produced into a distal tooth fl anked by 2 pairs of subdistal teeth, with a spinule in each interdental notch.

Distribution
Kara Sea, 207 m (Stappers 1911, as H. elegans); Svalbard, South of Jan Mayen, North of Faeroe Islands, Western Iceland, 700 to 1384-1435 m, +0.4°C to +2.0°C (Stephensen 1931(Stephensen , 1938, as H. elegans), as H. elegans); Northern Iceland, 407-996 m, -0.6°C to +0.1°C, i.e. in the Arctic Shallow Water Mass and the Arctic Bottom Water Mass (Weisshappel 2001, as H. elegans); Kara Sea, Laptev Sea, East Siberian Sea (Sirenko 2001).Brandt (1997) also records some 'cf.Halirages elegans' from East Greenland between 260 and 2681 m, which possibly include specimens of the present species.Remarks Stappers (1911) provided excellent fi gures of many parts of the holotype and Stephensen (1931) gave further good drawings of the species, so that there is no need for a completely new set of fi gures, which would require further destructive dissections of the single type specimen.The illustrations of Stappers (1911) are simply reproduced herein, with a few complementary original fi gures made without dissection.It can be observed that the left rami of uropod 1, as illustrated by Stappers (1911) is shorter than the right rami examined during the present study.The right rami were somewhat distorted and this was possibly also the case of the left rami.So, it is possible that Stappers (1911) made an imperfect reconstruction of the appendage.Kamenskaya, 1980 * The pleon of the type specimens of H. gorbunovi (and only known specimens) is missing and its ornamentation is therefore unknown.However, since that species looks very similar to H. qvadridentatus, it is assumed that some of the posterior body segments have a posterodorsal tooth.

Checklist of Halirages species
Halirages caecus Kamenskaya, 1980 The original spelling Halirages caecum, given by Kamenskaya (1980), has to be changed into H. caecus because Halirages is masculine and caecus, -a, -um is an adjective.Arctic Ocean, 'Drifting Station North Pole 22', stn 10, 81°44'N 121°46'W, 3467 m; stn 14, 81°42'N 126°51'W, 3530 m; stn 29, coordinates missing, 2910-2630 m; stn 34, 79°02'N 127°39'W (including holotype), 3290 m; stn 50, 78°45'N 127°36'W, 2990-2950 m; stn 53, 78°10'N 130°00'W, 2810 m.The table of stations provided by Kamenskaya (1980) indicates that all the stations of the drifting station North Pole 22 are in East longitudes.However, this is clearly a mistake and all these stations were actually in West longitudes.Indeed, Kamenskaya (1980) indicates that her stations are located in the Canadian Basin of the Arctic Ocean, which only fi ts with West longitudes.Kamenskaya (2001) refers both to positions equivalent to West longitudes (with an unusual coding system) and the Canadian Basin for the same stations.Finally the stations of the drifting station North Pole 22 are plotted on fi g. 2 of Afanas'ev and Filatova (1980) and their positions are indeed in the Canadian Basin, which confi rms they are in West longitudes.Stephensen, 1917[≠ H. bispinosus (Spence Bate, 1857) in Boeck 1871], type-locality: Bredefjord, South Greenland (Stephensen 1917).A very frequently recorded Arctic and subarctic species with a probably circumpolar distribution, reaching southwards South Iceland and the Skagerrak (Stephensen 1938), and occurring between 5 and 670 m (Oldevig 1959).On the fi gures of G.O. Sars (1893), which are otherwise excellent, the posterior border of the basis of P5-P7 is illustrated as totally smooth, whilst it is actually very weakly crenulate.In this respect, the fi gures of Stephensen (1917, as H. bispinosus) are more accurate.Gurjanova, 1946 According to Gurjanova (1946Gurjanova ( , 1951)), the two type specimens (6.5 mm long juveniles missing pleon) were collected at 2500 m depth in the Arctic Ocean, at Station 100 of the Sedov expedition.The coordinates of the station (not provided by Gurjanova 1946Gurjanova , 1951) ) are: "stn 100, 8-10 July 1938.RV 'Sadko'.81°10'N; 137°17'E; depth 2500 m; sediment, catch conditions and size not known; catch not rich, but satisfactory" (Gorbunov 1946).The station is located in the eastern part of the Nansen Basin.
Halirages stappersi sp.nov.Kara Sea, Laptev Sea, East Siberian Sea, Svalbard, South of Jan Mayen, North of Faeroe Islands, Northern and Western Iceland, 207 to 1384-1435 m (see descriptive account).

General discussion
A complete revision of the genus Halirages would have been desirable, but, unfortunately time did not allow for this, leaving pending questions.Small, unidentifi ed, defective Halirages specimens were found in recently collected samples from the deep Norwegian Sea.They could either indicate the occurrence of a further species in the area or they could be juveniles of a known taxon.The extensive latitudinal and longitudinal range (circum-arctic / circum-subarctic) of H. fulvocinctus raises the question of whether it is a genetically homogeneous species or not.Whilst H. nilssoni is a very distinctive species, it was only illustrated by old-fashioned fi gures (Ohlin 1895), which do no longer meet the required standards.Therefore, better illustrations of that species would be desirable (as for H. caecus and H. gorbunovi).Finally, and this point is essential, the delimitation of the genus Halirages is problematic.Due to the restricted scope of the present paper, the previously accepted composition of the genus has been retained herein.However, it appears that the borderline between Halirages Boeck, 1871, Apherusa Walker, 1891, and to a lesser extent Rozinante Stebbing, 1894 is blurred, an issue entirely disregarded in the revision of Apherusa by Krapp-Schickel & Sorbe (2005).With the exception of the alleged presence (Halirages) or absence (Apherusa) of calceoli in terminal males, a careful comparison of fi gures of species the two genera does not reveal any stabile differential character states.Character states previously considered as differential as the size and shape of article 3 of mandibular palp, the length of uropods and telson proved to be too variable to be retained.So, taken alone, the signifi cance of the presence/absence of calceoli seems highly dubious for separating the two genera, as Bousfi eld & Hendrycks (1997: 5) remark that "single-character diagnoses of higher taxonomic groups are inherently risky and unstable".Cladistic analyses of northern calliopiids could possibly reveal the existence of homogeneous clusters of species and call attention to synapomorphies suitable for circumscribing genera more reliably.
Canto XXXII, verse 58): fi rst round of the ninth circle of Dante's Inferno, which the poet describes as a frozen lake.The name alludes to the deep basin of the Norwegian Sea, which is the habitat of the species.With its negative temperatures, this body of icy abyssal water, trapped under a layer of warmer Atlantic waters is not unlike the frozen lake of Dante's Inferno.The vernacular noun in medieval Italian is Latinized as caina, -ae and is a genitive.

Table 1 .
(Tomczak & Godfrey 2003)orway, 2589-2615 m.Besides the type series listed above, several other specimens of H. cainae sp.nov.from the Norwegian deep Sea were found during the workshop 'Deepwater amphipods of the Norwegian Sea, Skibotn Feltstasjon, August 2nd-9th, 2009', and were at that time provisionally identifi ed as H. caecusKamenskaya, 1980.These specimens, which were not re-examined in this study, should have been deposited in the Zoological Museum of Bergen.In addition to the collected specimens, large red Halirages, which were presumably also H. cainae sp.nov., were observed on muddy bottoms with the video platform Campod (see e.g.Anonym 2006; Buhl-Mortensen et al. 2009 for description) during the MAREANO 2009-111 cruise.H. cainae sp.nov. is a species associated with the deep-sea Arctic water mass with negative temperature of the Norwegian Sea(Tomczak & Godfrey 2003).Like almost all other organisms collected in that sea during the second leg of the MAREANO 2009-111 cruise, they were already dead when arriving on deck, after crossing the warmer upper water mass of Atlantic origin.The ventral border of coxa 1 and 2 bears about 10 weak crenulations in H. cainae sp.nov.vs.about 20 pronounced serrations in H. qvadridentatus.H. cainae sp.nov.has a posterodorsal tooth on the seventh pereionite and the fi rst and second pleonite, never on the sixth pereionite (the dorsal dentition was also checked in the specimens seen during the workshop 'Deepwater amphipods of the Norwegian Sea').In adult and subadult H. qvadridentatus, there is always a tooth on the seventh pleonite, and usually also a tooth on the sixth pereionite.On the posterior margin of the basis of pereiopods 5-7, the number of crenulations or serrations is lower in H. cainae sp.nov.(9,14,18)than in H. qvadridentatus(17,  >22, 22-37).The posteroventral angle of the basis of pereiopod 7 is less angular in H. cainae sp.nov.than in H. qvadridentatus.The carpus and merus of pereiopod 7 are longer and more slender in H. cainae sp.nov.than in H. qvadridentatus.The carpus is 11.5 x as long as wide and 1.6 x as long as posterior border of basis in H. cainae sp.nov., whilst these ratios are 8.8 and 1.3 in H. qvadridentatus, respectively.The propodus is 18.4 x as long as wide and 1.8 x as long as posterior border of basis in H. cainae sp.nov., whilst these ratios are 13.0 and 1.4 in H. qvadridentatus, respectively.Finally, the tip of the telson has a single tooth in H. cainae sp.nov., whilst it is tridentate in H. qvadridentatus.Character states of Halirages species of the group qvadridentatus.inH.gorbunovi.In H. cainae sp.nov.theposteroventral angle of the basis of pereiopod 6 is rounded, whilst it is distinctly angular in H. gorbunovi (seeGurjanova 1946: 288 fi g. 21.4).In H. cainae sp.nov.. the posterior border of the basis of pereiopod 6 has about 9 crenulations vs. about 5 scarcely distinct crenulations in H. gorbunovi.The Russian text also indicates that the basis of pereiopods 5 and 7 have an almost smooth posterior border, while H. cainae sp.nov.has small, but distinct serrations on the posterior border of the basis of pereiopod 7. Finally, H. cainae sp.nov.has eyes (which disappear only after a long preservation period), whilst H. gorbunovi is said to have none.This absence of eye will have to be confi rmed when fresh specimens of H. gorbunovi will be available to study.It must be noted that the pleon of the two syntypes of H. gorbunovi is missing and that they are small juveniles: 6.5 mm from the tip of the rostrum to the end of the pereion (the largest H. cainae sp.nov. is 40 mm long).An adequate characterization of H. gorbunovi would only be possible when topotypical specimens from a wide size range will be available for study.
Stephensen (1931)cainae sp.nov.belongs to the Halirages species of the group qvadridentatus, together with H. qvadridentatus G.O.Sars, 1877, H. gorbunovi Gurjanova, 1946 and H. caecusKamenskaya, 1979.The accounts given by G.O.Sars (1885)andStephensen (1931)provide a clear picture of the diagnostic characters of the North-European species H. qvadridentatus, while the descriptions of the two other species (from the Arctic Ocean) are very defi cient.The differences between the four species are described in the next section and are summarized in table 1.H. cainae sp.nov.canbeseparatedfromH. qvadridentatus by the following characters.The eye is small, reniform to subreniform in H. cainae sp.nov., whilst it is large and quadrato-elliptic in H. qvadridentatus (see G.O. Sars 1885, plate 14 fi g. 4).In H. cainae sp.nov., the junction between the anterior and ventral borders of coxa 1 forms a square angle, while in H. qvadridentatus it forms a tooth pointing forwards.H. gorbunovi is very inadequately described but exhibits the following differences with the present species.In H. cainae sp.nov.thecrenulations of the ventral border of coxa 1 and coxa 2 are weaker thanThe description of H. caecus is also very defi cient but includes the following differences with the new species.In H. cainae sp.nov., article 2 of mandibular palp is stout and is densely setose, all along its medial margin, whilst it is slender and sparsely setose in H. caecus.In H. cainae sp.nov., article 3 of mandibular palp is apically broader than in H. caecus.In H. cainae sp.nov., the anteroventral corner of coxa 1 forms a square angle, whilst in H. caecus it forms a tooth pointing forward.It must be pointed out that pereiopods 5-7 of the types of H. caecus were neither illustrated nor described and were possibly missing in all specimens.Finally, H. caecus is said to be eyeless whilst H. cainae sp.nov.doeshaveeyes(which can become indistinguishable after a long preservation period).As for H. gorbunovi, the absence of eyes needs confi rmation.Halirages qvadridentatus -G.O.Sars 1885: 172, pl.14 fi g. 4. -d'Udekem d'Acoz 2010: 146 (discussion on spelling).Halirages quadrispinosus -G.O.Sars 1893: 436 (lapsus calami).Acanthozone quadridentata -Della Valle 1893: 611 (pro parte), pl.59 fi g. 22. Islands, RV Knight Errant, stn 8: anterior part of head, right Gn1, right Gn2, left P7, left Ep3, left U3, 2 telsons, BM(NH), slide 1911.11.8.1269; RV Knight Errant, 1880, stn 8, label of slides indicating "main organs (mandible from three specimens)": 3 mandibles, 1 upper lip, 1 maxilliped, in poor condition, BM(NH), slide 1911(NH), slide  .11.8.1270.  .