New insights into the diversity and ecology of benthic hydroids (Cnidaria, Hydrozoa) from the Ross Sea (Antarctica)

Hydrozoans are recognized as one of the main and most characteristic zoological groups of the Antarctic benthos, despite the fact that there are still large Antarctic areas where the hydrozoan fauna is completely unknown or poorly known (e.g., the Admunsen Sea and Mary Byrd Land, in West Antarctica, and Queen Maud Land and Enderby Land, in East Antarctica). The present study contributes to a better understanding of the Ross Sea benthic hydroid fauna by studying material collected through several New Zealand expeditions mostly with RV Tangaroa. The Ross Sea includes the world’s largest marine-protected area (MPA) and is of considerable biological value and importance for scientific research. Although some parts of the Ross Sea shelf have been intensively sampled, others have not, including deeper parts of the continental shelf and the slope. Forty species were found, belonging to 15 families and 19 genera. Six species, including Eudendrium megaloarmatus sp. nov., Nemertesia gelida sp. nov., Schizotricha frigida sp. nov., Symplectoscyphus pseudofrondosus sp. nov. and Symplectoscyphus tortuosus sp. nov., represent new records, bringing the number of known species in the Ross Sea to 84. Leptothecata is dominant, with 35 species, while Anthoathecata is represented by five species. Symplectoscyphidae is the most diversified family with 12 species (30%), and Symplectoscyphus is the most speciose genus with nine species (23%). Three main hydroid assemblages have been found in the studied area, two with a wide bathymetric range and relatively high species diversity, and a third with a narrow and deep bathymetric range and remarkably low hydroid diversity. The hydroid fauna is dominated by species with a wide bathymetric distribution and virtually all species are restricted to Antarctic or Antarctic/sub-Antarctic waters, with 70% endemic to the Antarctic region.


Introduction
The Ross Sea region contains the largest marine-protected area (MPA) in the world, declared in 2017 with the recognition of a requirement for scientific evaluation of its ecological status, spatial adequacy, and effectiveness (O'Driscoll et al 2019). Research is therefore a fundamental feature of this MPA.
Even when benthic hydroids are known from only a few areas of the Ross Sea, particularly off the west coast (e.g., Cape Adare, Cape Hallet, McMurdo Sound), the region is acknowledged as one of the most diverse Antarctic areas for this zoological group (Peña Cantero 2017), with 78 species known so far to inhabit this zone (Peña Cantero 2019). Benthic hydroids from the Ross Sea were reviewed by Peña Cantero (2017), after studying a collection gathered during the New Zealand BioRoss 2004 survey. Peña  study, based on hydroids collected during the New Zealand TAN0802 voyage, was particularly important because the bathymetric range investigated extended to the continental slope, thus including depths little explored for Antarctic benthic hydroids; knowledge about the group in the Ross Sea originated mainly from shelf depths (cf. Peña . As a result, the depth range of 15 species was increased, notably in most cases, and, thus, several species that had been considered shelf species were demonstrated to inhabit the slope bottoms as well.
I present here the results of the study of a collection of benthic hydroids from the Ross Sea collected during several New Zealand research surveys. The aim of the study is to 1 3 increase the available knowledge on the biodiversity, ecology, and distribution of benthic hydroids inhabiting this vital Antarctic area and thus contribute to the baseline information about the established MPA to allow scientific evaluation and monitoring.

Material and methods
Samples studied come from the Ross Sea and were mostly collected during several New Zealand research surveys with RV Tangaroa. Most stations are located at the outer edge of the continental shelf and upper slope (see Online Resource 1). Data associated with the samples containing hydroids are given in Online Resource 2. Different sampling gears (Van Veen grab, epibenthic sled, bottom trawls, bottom longline) were employed. Twenty-four samples containing hydroids were collected at depths between 49 and 1865 m. Hydrozoans were preserved in 70% ethanol.
The material studied is deposited at the National Institute of Water and Atmospheric Research (NIWA) Invertebrate collection in Wellington, New Zealand. The material examined for each species is described in Online Resource 3.
The collection primarily comprises relatively well-known species. Consequently, no descriptions are provided, except for those species new to science. Likewise, as information on the ecology and distribution of most species has been covered in recent publications, the discussion will focus on the new data obtained in the present study.
The geographical and bathymetric distribution has been studied using the groups recognized by Peña Cantero and García Carrascosa (1999) and Peña , respectively.
The colonial nature of most hydrozoans and the fact that colonies often come on board fragmented or in poor condition make quantitative studies based on traditional sampling gears difficult, if not impossible (Peña Cantero 2021). The species found were divided into four groups according to their frequency of occurrence (Peña Cantero and Majón-Cabeza 2014) as a proxy of their importance in the benthic communities: ubiquitous (≥ 30% of stations), very common (between 30 and 20%), common (between 20 and 10%), and rare or accidental species (≤ 10%).
To investigate benthic hydroid assemblages, a presence/ absence dataset was created, and Sørensen's similarity index was used to generate a similarity matrix, following the methodology employed by Soto Àngel and Peña . Agglomerative hierarchical clustering using the group average method was performed on the similarity matrix, accompanied by the SIMPROF test to establish the statistically supported relationships between samples and to investigate the hydroid assemblage structure. Additionally, a SIMPER analysis was conducted to identify the key species characterizing each assemblage. All analyses were performed using the PRIMER 6 software package (v.6.1.6) (Clarke and Warwick 2001).

Diversity
Forty species were found in the NIWA hydroid collection gathered during several New Zealand campaigns carried out in the Ross Sea area (Table 1); three species could only be identified at species level. Six species represent new records for the area, bringing the number of known species in the Ross Sea to 84.
In general, most of the species found were present in only a few stations. In fact, 30 species were found only in one or two stations. Symplectoscyphus frondosus and Schizotricha nana are the species with the highest occurrence, being present in five stations each (20.8%), followed by Billardia subrufa and Symplectoscyphus exochus, in four (16.7%), and Abietinella operculata, Halecium incertus, Oswaldella terranovae, Staurotheca nonscripta, Symplectoscyphus glacialis, and Symplectoscyphus liouvillei in three (12.5%).

Hydroid assemblages
Three distinct hydroid assemblages have been obtained from the cluster analysis ( Fig. 1), all of them are supported by the SIMPROF test (p 0.05). Assemblage A comprises eight stations, at depths from 198 to 1124 m; B comprises four, at depths between 733 and 866 m; and C includes six stations, at depths from 49 to 1100 m. Six stations were not grouped in the analysis.
The study of the species related to each group has made it possible to determine the hydrozoan fauna characterizing them (Online Resource 4). Assemblage A, with a similarity of 18.00, is defined by B. subrufa, H. incertus, S. nana and S. nonscripta. Group B, with a similarity of 83.33, is characterized by S. frondosus, which is the only species accountable for all similarity. Finally, assemblage C, with a similarity of 17.69, is mainly defined by A. operculata, S. exochus, S. liouvillei, and O. terranovae.
The hydroid assemblages retrieved show a strong dissimilarity between them. The lowest dissimilarity (95.39) is found between groups B and C, closely followed by that between A and C (96.38); the highest dissimilarity (100) is observed between A and B. Description. Basally broken stem 130 mm high, strongly and completely polysiphonic, very robust, about 10 mm in diameter at basal part. Stem irregularly branched; up to fourth-order branches present. Some anastomoses present.
Stem consisting of a bundle of stolons, covered with smooth, firm perisarc, running alongside each other and irregularly giving rise to short pedicels bearing a single distal polyp. Perisarc pedicel ending well below base of polyp, short and completely annulated on distal branches, longer and only basally ringed on more basal parts.
Hydranths relatively large, urn-shaped, provided with a pedunculate, trumpet-shaped or spherical hypostome and a whorl of about 25 filiform tentacles.
Gonosome unknown. Remarks. Eudendrium megaloarmatus sp. nov. is characterized by its unique colony structure. None of the numerous described species of Eudendrium has a colonial structure like the one present in this species. Whereas there are polysiphonic species in Eudendrium, the structure of the colony is completely different, with polyps originating from a single main tube, which usually re-branches in varying degrees. The remaining tubes are typically simple accessory stolons, with a support function, which do not give rise to polyps. It is known that in Eudendrium glomeratum Picard, 1952 secondary tubes of polysiphonic colonies can also give rise to branches (Peña Cantero and García Carrascosa 2002). Nevertheless, in this case, a main stem is still clearly identifiable and those accessory tubes form branches that re-branch. It would be interesting to re-examine that material to see if all accompanying stolons give rise to branches and to assess the possibility that the stolons giving rise to branches are actually from other planulae. In any case, as shown above, in Eudendrium megaloarmatus sp. nov., there is no main tube, instead all the stolons conforming the stem give rise to small, unbranched pedicels provided with distal polyps. Ecology and distribution. Eudendrium megaloarmatus sp. nov. was collected at a depth of 412-420 m in Moubray Bay, Ross Sea.
Etymology. The specific name megaloarmatus makes reference to the very large microbasic mastigophores present in this species.
Description. Stem 95 mm high, basally truncated, distinctly tortuous, stiff, and strongly polysiphonic throughout its whole length, diameter at base about one millimeter. Stem giving rise to primary branches spirally arranged; a complete whorl every centimeter and four primary branches. Primary branches polysiphonic for much of their length, directed upwards, but arching downwards later, and giving rise to lower-order branches alternately in two planes (up to fourth-order branches observed). The number of hydrothecae between every two branches is uneven. Branches divided into internodes delimited by alternating oblique nodes. Internode length decreasing distally (e.g., from 1100 µm at first internode after branching to 600 µm at distal end of branch). Internodes giving rise to new branches markedly bent.
Hydrothecae on distal part of internodes, arranged alternately, usually not reaching subsequent alternate ones. Hydrotheca cylindrical, slightly abcaudally directed. Maximum diameter at point adcauline wall becomes free, decreasing slightly toward aperture and more sharply to base. Adcauline hydrothecal wall adnate to internode for less than half its length. Free part of adcauline hydrothecal wall convex; adnate part almost straight. Abcauline hydrothecal wall slightly concave by the middle. Cusps of hydrothecal aperture blunt, separated by shallow embayments. Rim of hydrothecal aperture usually with renovations.
Remarks. The primary branches usually give rise to three secondary branches (the first two alternate in an open angle). Typically, the first secondary branch is distinctly long, bends downwards as the primary branch, and gives rise to two third-order branches. The other two secondary branches are shorter and do not branch again. On one occasion, on the fifth primary branch from the top, there were four secondary branches, with the first three partially polysiphonic; the second one had given rise to three third-order branches, and the first of these to three short fourth-order branches.
Symplectoscyphus pseudofrondosus sp. nov. is morphologically close to S. frondosus in the general colony structure, but they differ in several important features. Both share a stiff, strongly tortuous and polysiphonic stem giving rise to spirally arranged primary branches. However, colonies of S. frondosus are much more robust and dense, with all branches tightly packed and of similar development, giving the colonies a bottlebrush appearance (see Fig. 1C, in Peña Cantero 2010, for a colony of similar height). In Symplectoscyphus pseudofrondosus sp. nov., primary branches are less rigid and more loosely packed, and their length gradually increases distally. Even though primary branches are spirally arranged in both species, a whorl is completed every centimeter and four primary branches (i.e., there are three primary branches between two with the same orientation) in Symplectoscyphus pseudofrondosus sp. nov., whereas in S. frondosus it takes about 5 mm and five branches, which indicates the higher density of primary branches. They also differ in the hydrothecal arrangement: densely packed in S. frondosus, with the distal part of the hydrothecae clearly overlapping the basal part of the following alternate ones. In addition, they differ in the distinctly larger size of the hydrotheca of Symplectoscyphus pseudofrondosus sp. nov., which also has a distinctly longer free adcauline wall. Finally, the hydrotheca is straight or just slightly abcaudally directed in S. frondosus, with the cusps of the hydrothecal aperture separated by deeper embayments.
Ecology and distr ibution. Symplectoscyphus pseudofrondosus sp. nov. was collected at a depth of 815-821 m in Iselin Bank.
Etymology. The specific name pseudofrondosus refers to the fact that the shape of the colony in this species is reminiscent of that of S. frondosus. sp. nov. (Figs. 4b,7) Material examined. TAN 1502/55, one stem 90 mm high (Holotype, NIWA 104504).

Symplectoscyphus tortuosus
Description. Stem 90 mm high, basally broken, distinctly tortuous and polysiphonic, except for the last 10 mm. Most distal polysiphonic part with a single accompanying stolon; number of stolons increasing basally and thus basal part becoming strongly polysiphonic. Diameter at base c. 800 µm. Branching usually alternate in two planes, making an acute angle. Branches originating almost perpendicularly to plane formed by hydrothecae of preceding branches. Primary branches basally polysiphonic; one of them much more developed, becoming a lower-order stem. Branching up to fifth order. The number of hydrothecae between every two branches is uneven. Branches divided into internodes delimited by alternating oblique nodes. Internode length decreasing distally (e.g., from 950 µm at first internode after branching to 550 µm at branch distal end). Internodes giving rise to new branches markedly bent.
Remarks. Branching at the distal, highly branched part is in general alternate in two planes making an acute angle. The pattern is less clear in the much less branched basal part, where some branches are lost, and their origin is hidden by stolons. However, some alternate branches are also visible. The consequence of this form of branching is that the branches tend to be situated on one side of the stem. As a result, an anterior and posterior part of the colony are distinguishable, with most branches on the frontal side. On the other hand, branching occurs at uneven intervals, but at every fifth hydrotheca is the most common, followed by branching at every third hydrotheca.
Symplectoscyphus tortuosus sp. nov. is morphologically closer to S. frondosus and S. liouvillei. Although the general appearance of the colony and the shape of the hydrotheca are similar to those of S. liouvillei, there are important differences that allow separating both species. The stem in S. liouvillei is sinuous (see Fig. 14, in Billard 1914, for the type material, a colony of similar height, and Fig. 1D in Peña Cantero 2010), but not as winding as in Symplectoscyphus tortuosus sp. nov. The branching pattern is also different, as the stem gives rise to primary branches following a spiral pattern in S. liouvillei (Billard 1914;Peña Cantero 2010), but it is alternate in two planes in Symplectoscyphus tortuosus sp. nov. The bourrelet at the base of the internode, so characteristic of Billard's species, is not present is Symplectoscyphus tortuosus sp. nov. Although the general shape of the hydrotheca is similar, in the present species, the free part of the adcauline wall is distinctly longer, and the adnate part is shorter than in S. liouvillei. Billard's species is characterized by hydrothecae with the adnate adcauline wall much longer than the free part (two-thirds of the adcauline wall are adnate according to Billard 1914), contrary to what happens in Symplectoscyphus tortuous sp. nov. The hydrotheca in S. liouvillei is also wider than that of Symplectoscyphus tortuous sp. nov. (e.g., diameter at aperture, maximum diameter, and diameter at diaphragm are 200-220, 230-250 and 170-200 µm, respectively, in S. liouvillei in Peña Cantero 2017). The cusps in S. liouvillei are also sharper and separated by deeper embayments. Finally, in S. liouvillei, the distal part of the hydrothecae usually reach the basal part of the following alternate ones. In the present species, however, the hydrothecae typically do not reach the successive alternate ones (Fig. 7c), although they may do so in some parts (Fig. 7b).
Symplectoscyphus tortuosus sp. nov. clearly differs from S. frondosus in several important features. Symplectoscyphus frondosus has stiff, tightly branched, bottlebrush colonies with all branches of similar development, whereas the colony is less firm, less branched, and with unevenly developed, droopy branches in Symplectoscyphus tortuosus sp. nov. They also differ in the polysiphonic development, as polysiphony is strongly developed over the whole colony, including the branches, in S. frondosus (only the most distal part of the branches is usually monosiphonic). Symplectoscyphus frondosus has stems with primary branches arranged in a spiral fashion (Peña Cantero 2010), whereas primary branches are alternate in two planes in Symplectoscyphus tortuosus sp. nov. In addition, they differ in the distinctly larger size of the hydrotheca of Symplectoscyphus tortuosus sp. nov., which also has a distinctly longer free adcauline wall. Furthermore, in S. frondosus, the hydrotheca is straight or just slightly abcaudally directed, and the cusps of the hydrothecal aperture are separated by deeper embayments.
Symplectoscyphus tortuosus sp. nov. also resembles Symplectoscyphus pseudofrondosus sp. nov. in the general colony appearance and the hydrothecae are of similar shape and size. However, the details of the colony structure are completely different. In Symplectoscyphus pseudofrondosus sp. nov., the stem is polysiphonic over its whole length and the primary branches are polysiphonic for much of its extension, but in Symplectoscyphus tortuosus sp. nov. a significant distal portion of the stem is monosiphonic and only the basal part of the primary branches is polysiphonic. In addition, in Symplectoscyphus pseudofrondosus sp. nov., the primary branches are spirally arranged and there is no secondary stem. Finally, the free part of the adcauline wall is slightly longer and the adnate part shorter in Symplectoscyphus pseudofrondosus sp. nov.
Ecology and distribution. Symplectoscyphus tortuosus sp. nov. was collected at depths between 1753 and 1865 m in the Ross Sea Central Basin.
Etymology. The specific name tortuosus has been taken from the Latin adjective 'tortuosus' and refers to the winding, tortuous stem.
Description. Stems polysiphonic, branched, at least 160 mm high, consisting of a main axial tube and several accompanying ones, whose number decreases distally; most distal part of stem only with main tube. Primary axial tube on one side of stem, not covered by accessory tubes for most of its length. Main tube divided into homomerous hydrothecate internodes bearing hydrothecae and nematothecae. Secondary tubes provided only with nematothecae.
Cauline internodes with very long apophysis provided with two nematothecae. Cauline apophyses arranged alternately in two planes, forming an obtuse angle. Cauline internodes with a hydrotheca at axil between apophysis and internode and four nematothecae: two flanking hydrothecal aperture and two infrahydrothecal nematothecae. Cauline apophyses forming an acute angle with stem and giving rise to hydrocladia.
Hydrocladia divided into internodes and branched; up to fifth-order hydrocladia present. Lower-order hydrocladia originating from first internode of successive hydrocladia. Hydrocladial apophyses very long, forming an acute angle with internode, and provided with one or two nematothecae. Length and diameter of hydrocladial internodes roughly constant, decreasing only slightly along hydrocladia.
Unforked hydrothecate hydrocladial internodes with one hydrotheca on its distal half and three nematothecae: two flanking hydrothecal aperture and one mesial infrahydrothecal nematotheca.
Hydrotheca elongate, length increasing along hydrocladia (e.g., 250 µm at 1st unforked hydrocladial internode and 390 at 13th one). Adcauline wall completely adnate to internode. Abcauline wall either straight or slightly convex in basal Gonothecae inserted on small apophyses placed just below hydrotheca of forked and unforked hydrocladial internodes. Male gonothecae elongate pear-shaped, with a slightly oblique circular aperture occupying almost the whole distal flat part, and a basal chamber delimited by a circular diaphragm situated around basal sixth; usually one nematotheca (two also observed) on adcauline side of basal chamber. Female gonotheca fusiform, with kidney-shaped subterminal aperture and a basal chamber delimited by a circular diaphragm around basal fifth; two nematothecae on adcauline side of basal chamber and one on abcauline side.
Remarks. The branched nature of the colonies of Schizotricha frigida sp. nov. is clearly seen in the available material: a distal colony fragment, 100 mm long, is provided with seven secondary stems, which do not re-branch; another stem fragment, 115 mm long, has at least four lower-order stems at different level; and the most basal stem fragment, 135 mm long, has at least three lower-order stems also originating at different level.
From the known Antarctic/sub-Antarctic species of the genus, Schizotricha frigida sp. nov. is morphologically closer to Schizotricha turqueti Billard, 1906 andSchizotricha glacialis (Hickson andGravely, 1907) by the presence of a single infrahydrothecal nematotheca on the unforked hydrocladial internodes, the absence of ahydrothecate internode following the cauline and hydrocladial apophyses and the elongate hydrotheca. They have, however, important differences. Schizotricha frigida sp. nov. mainly differs from both species by their unbranched stems. As noted above, despite the fragmented condition of the available material, Schizotricha frigida sp. nov. is undoubtedly characterized by having branched stems. Its branching pattern is similar to that present in species such as Schizotricha unifurcata Allman, 1883 or Schizotricha trinematotheca Peña Cantero and Vervoort 2005, but different from that present in species such as Schizotricha anderssoni Jäderholm, 1904 or Schizotricha nana Peña Cantero et al. 1996, where lowerorder stems do originate from a further development of the hydrocladia. In S. anderssoni, for example, according to Peña Cantero and Vervoort (2005 p. 801), 'there is continuous branching of the stem. … this species usually lacks branched hydrocladia. When branched, hydrocladia are either merely bifurcated or they become lower-order stems'; similar branching is present in S. nana (cf. Peña Cantero et al. 1996). The branching in Schizotricha frigida sp. nov., however, as indicated above, is similar to that of S. trinematotheca in which new stems seemingly originate from accessory tubes of the polysiphonic part of older stems (Peña Cantero 2019). Schizotricha glacialis and S. turqueti are characterized by unbranched stems, although secondary stems have occasionally been observed in the latter, clearly as consequence of regeneration after fracture of the original stem (Peña Cantero and Vervoort 2005;, 2017. They also differ in that the cauline internodes are consistently provided with two infrahydrothecal nematothecae in Schizotricha frigida sp. nov. but have one or two in S. glacialis and S. turqueti. In addition, the cauline apophyses are much longer in Schizotricha frigida sp. nov. (600 to 700 µm) than in S. turqueti and S. glacialis (328 µm and 328-361 µm, respectively, in Peña Cantero et al. 1996) and are provided with two nematothecae, whereas only one nematotheca has been described in S. glacialis and there is usually one, but two have also been observed in S. turqueti.
Ecology and distribution. Schizotricha frigida sp. nov. was collected at a depth of 49 m off Cape Hallett. Gonothecae in February.
Etymology. The specific name frigida comes from Latin adjective 'frigidus' and refers to the cold conditions of the habitat where this species lives.
Description. A distal stem fragment, 35 mm long, polysiphonic except for the last four millimeters. Hydrocladia originating from cauline apophyses arranged in alternate verticils of three, giving rise to six longitudinal rows. Stem divided into internodes by transverse nodes, usually with one verticil of three apophyses at the same level, but internodes with two verticils also present. Hydrocladia with up to seven hydrothecae. Cauline apophyses with one mamelon, two axial nematothecae and two above mamelon.
Stem with two short, monosiphonic secondary stems originating from an accompanying stolon on the same side of stem and almost completely aligned: more distal one, 5 mm long, at nine millimeters from distal end of stem; more basal secondary stem, 8 mm long, originating 5 mm below. Second-order stems resting on a long apophysis, with one nematotheca, and beginning with a short ahydrothecate internode, also provided with one nematotheca, followed by an internode with one apophysis and a second internode with two almost opposite apophyses, situated at the same height and practically perpendicular to the lower apophysis. Successive internodes already provided with verticils of three apophyses.
Hydrocladia divided into internodes, alternately hydrothecate and ahydrothecate, starting with a hydrothecate internode. Ahydrothecate internode with a very basal nematotheca resting on a strong swelling of internode.
Hydrothecate internode with a hydrotheca resting approximately in the middle of internode and three nematothecae: two flanking hydrothecal aperture and one mesial infrahydrothecal nematotheca not reaching bottom of hydrotheca. Internode with strongly developed perisarc ridges: one before mesial nematotheca, one in the middle between mesial nematotheca and hydrotheca, one at base of hydrotheca, one just above hydrothecal aperture and a fifth before distal node. Ridges becoming less conspicuous distally along hydrocladia.
Hydrotheca low, roughly as high as wide. Hydrothecal aperture circular, perpendicular to longitudinal axis of internode; rim even. Nematothecae bithalamic; distal chamber with scooped rim. Paired nematothecae of hydrothecate internodes wider and much longer than infrahydrothecal nematotheca, with markedly widening distal chamber.
Measurements ( Remarks. Peña Cantero (2008) first recorded the genus and the family for Antarctic waters, specifically off Deception Island and the south of Livingston Island. The infertile and poor condition of his material precluded its identification at the species level, although the specimens were morphologically similar to Nemertesia antennina (Linnaeus, 1758). This study represents the second record of Nemertesia in the Antarctic Region, as well as the first record of a second species of the genus, as the present material clearly differs from Peña Cantero's (2008), which consisted of unbranched and monosiphonic stems. Consequently, the present study describes for the first time a species of Nemertesia and the family Plumulariidae for Antarctic waters.
The material described here is scarce and many characters that could potentially be variable are difficult to evaluate. However, many taxonomic and biogeographic reasons remain to consider the material examined as a new species.
Nemertesia gelida sp. nov. is morphologically close to Nemertesia cymodocea (Busk, 1851), although they are clearly different species. They share the polysiphonic, branched stems, and the structure of the hydrocladial apophyses, provided with a mamelon and two pairs of nematothecae, one at the axil of the apophyses with the stem and the other above the mamelon (according to Vervoort and Watson 2003, there is a pair of nematothecae in the axil and one nematotheca immediately above the mamelon). They differ because, in N. cymodocea, the cauline apophyses are arranged in decussate verticils of three or four apophyses, but also in decussate pairs in younger parts of the colony (cf. Ramil and Vervoort 2006); in the present species, there are only decussate verticils of three apophyses.
They also share the heteromerous division of hydrocladia into internodes with alternating hydrothecate and ahydrothecate internodes. However, they differ because, in N. cymodocea, the first hydrocladial internode is ahydrothecate, whereas it is hydrothecate in Nemertesia gelida sp. nov.
Both species have small, cup-shaped hydrothecae, with straight abcauline wall, but, in N. cymodocea, the rim is sinuous and the aperture curves toward the internode (Ramil Fig. 8 Schizotricha frigida sp. nov. a cauline internode showing hydrotheca, nematothecae, and cauline apophysis with two nematothecae; b forked hydrocladial internode showing hydrotheca, nematothecae, and hydrocladial apophysis with one nematotheca; c, d unforked hydrocladial internodes with hydrothecae and nematothecae; e female gonotheca; f male gonotheca. Scale bars: 250 µm ◂ and Vervoort 2006). In the present species, the rim is even, and the aperture is roughly perpendicular to the long axis of internode.
They also differ in the shape of the lateral nematothecae. In N. cymodocea they are elongated conical and similar to the infrahydrothecal nematotheca, whereas in Nemertesia gelida sp. nov., as stated above, their distal chamber markedly widens distally, and they are distinctly larger than the infrahydrothecal nematotheca.
According to Vervoort and Watson (2003), in N. cymodocea, the polysiphonic stem has long branches arising just above the base; and the branches re-branch basally into long upwardly directed shoots. In the present species, even when the available material is a single distal Finally, there are biogeographical reasons, as N. cymodocea seems to be restricted to temperate waters. According to Vervoort and Waston (2003), it is known from waters around southern Africa, New Zealand, and southwestern Atlantic. The present species, the first one described from Antarctic waters, is from the Ross Sea, in High Antarctica.
Nemertesia gelida sp. nov. is also morphologically similar to Nemertesia ciliata Bale, 1914. Both species share the presence of branched and polysiphonic stems and the heteromerous division of hydrocladia into internodes. However, as in the case of N. cymodocea, the hydrocladia start with an ahydrothecate internode, contrary to what happens in Nemertesia gelida sp. nov. In addition, in N. ciliata, the cauline apophyses are arranged in verticils of three or four, but also in decussate or even alternate pairs in younger parts (Ramil and Vervoort 2006). Furthermore, the hydrotheca is well below the middle of the internode and the rim of the hydrothecal aperture is oblique, curving inwards adcaudally. Finally, there are also biogeographical reasons as the geographically closer records of N. ciliata are from Tasmania, New Zealand and waters around South Africa.
Ecology and distribution. Nemertesia gelida sp. nov. was collected at a depth of 1086-1097 m, off Cape Adare.
Etymology. The specific name gelida comes from the Latin adjective 'gelidus' and refers to the ice-cold, frosty conditions of the Antarctic Ocean.

Biodiversity
As previously mentioned, in the collection studied, gathered through several scientific surveys in the Ross Sea, 40 hydroid species were identified. The diversity found here is in between the two most recent surveys in the area. Peña  reported 61 species from the Ross Sea, twenty-two representing new records for the area, thus raising the number of valid, known species to 77. Later, Peña Cantero (2019) accounted for 27 species from the shelf and slope of the Ross Sea, with a new record for the area, raising the number of known species to 78. In the present study, I report 40 species from the area, with six new records, H. incertus and five new species to science, thus raising the number of benthic hydroids known from the Ross Sea to 84. This figure makes this region third in terms of benthic hydrozoan diversity among the larger Antarctic areas, only behind the Antarctic Peninsula, with 112 species (cf. Mercado Casares et al. 2017), and the Weddell Sea, where 89 species have been reported (Soto Àngel and Peña Cantero 2019).
It is remarkable, as shown above, that only a few very speciose families and genera provide the majority of the species diversity in the collection studied. Only four families (Haleciidae, Schizotrichidae, Staurothecidae, and Symplectoscyphidae) and five genera (Antarctoscyphus, Halecium, Schizotricha, Staurotheca, and Symplectoscyphus) account for 57.5% of the species diversity. The remaining diversity is distributed among 12 families, most of which have only one species (two in the case of Campanulariidae, Campanulinidae, Kirchenpaueriidae, and Tubulariidae), and 14 genera, with the majority having a single species (only Campanularia and Tubularia are represented by two). This uneven distribution of the species diversity at the genus and family level is a key feature of the benthic hydroid fauna (Peña Cantero 2014a) that has been acknowledged in many other Antarctic areas (see, for example, Peña Cantero 2008Cantero , 2014b. Anthoathecata is represented by only five species in the present study, while Leptothecata is much more abundant with 35 species (12.5% and 87.5%, respectively). This result is in line with those recently found in the Ross Sea by Peña , who reported ten anthoathecates and 51 leptothecates (16,4 and 83,6%, respectively), and Peña Cantero (2019) who only found 27 leptothecates. Peña , however, found anthoathecates representing over 50% of the species present in a collection from Tethys Bay, also in the Ross Sea (11 out of 20 species studied). The difference could be correlated to the depths surveyed and the sampling methods. In Peña , the material was collected up to a depth of 48 m by scuba diving, whereas in the present study and those by Peña , the material, from much deeper bottoms (49-1753, 64-1014 and 283-2283 m, respectively), was collected with indirect sampling gears.

Substrate
There is a limited amount of information available regarding the substrate used by benthic hydroids when they are collected by indirect sampling methods. This scarcity is particularly notable for large species that are directly attached to the bottom, as they are often detached or basally broken on board (Peña Cantero 2021). The species that form smaller colonies provide more information, as many are found growing epibiotically on other organisms. Rhizorhagium antarcticum is the species observed on the highest number of substrates (algae, Antarctoscyphus spiralis, Lafoea dumosa, O. terranovae, S. nana, and S. glacialis), followed by Campanularia sp. (bryozoans, Antarctoscyphus elongatus, L. dumosa, O. terranovae and Symplectoscyphus naumovi). The picture is clearly distorted, since a collection of previously sorted hydroids certainly does not include all potential biological and non-biological substrates. As a result, the hydroids themselves provide almost all the information about the substrate (Peña Cantero 2021).
Concerning the species acting as basibionts, nine species were found to provide substrate for other hydroids; most of them, however, were hosting only one or two epibionts. In general, they are species that form small to medium-sized monosiphonic colonies that do not seem to provide a particularly suitable substrate. Conversely, there are species of hydroids that form large colonies that seem to constitute an appropriate substrate for species of hydroids typically found epibiotic on other organisms. The best examples in the collection are S. nana and O. terranovae, hosting nine and five hydroid species, respectively. The latter was found in three stations and S. nana in five. The role of these two species as basibionts of other hydrozoans has already been pointed out (Peña Cantero 2021), particularly in the case of O. terranovae, which has been considered a hotspot of epibiont biodiversity (Peña Cantero 2021), being home for 34 different species of hydrozoans. This, coupled with its high local abundance in some Antarctic areas (e.g., the Ross Sea, the Weddell Sea and Adélie Land), led it to be considered as a habitat builder (Peña Cantero 2021), in line with what has been described for other benthic hydroids as important constituents of animal forests (Di Camillo et al. 2017).
Alternatively, there are species forming large polysiphonic colonies that, however, are deprived of epibionts (e.g., Schizotricha frigida sp. nov., S. turqueti, S. liouvillei, Symplectocyphus pseudofrondosus sp. nov., and Symplectoscyphus tortuosus sp. nov.). This could be in part related to their low representation in the collection. However, S. frondosus is particularly remarkable because it is relatively abundant in the collection (it is present in five stations) but devoid of epibionts, except for a small stem of Tubularia. The present study is not an exception: Peña  found S. frondosus in five samples and neither colony had epibiotic hydroids and Peña Cantero (2019) reported it from ten stations and again none of the colonies had hydroids growing on them. This clearly indicates that this species is not suitable for the settlement and/or development of other hydroid species and points to the presence of some sort of antifouling mechanism. The species forms bottlebrush-shaped colonies, densely branched from the base, and polyps may protect the whole colony. The exception indicated above could be related to the fact that the polyp of Tubularia is gigantic, compared to those of S. frondosus, and is protected by a strong and relatively huge tube of perisarc.

Hydroid assemblages
The fact that about two-thirds of the stations have a very low diversity (1-2 species) is reflected in the results obtained in the analysis. The interpretation of the results is obviously constrained by this fact.
As mentioned above, three hydroid assemblages were found in the study area. The most widespread group, which includes eight stations, is assemblage A. The other two groups are less represented, with six (assemblage C) and four stations (assemblage B).
The highest internal similarity (83.33) is found in assemblage B, while it is distinctly lower, but similar, in A and C (18.00 and 17.69, respectively) (Online Resource 4). Assemblages A and C have, however, a high dissimilarity between them (96.38), as well as with B (100 and 95.39, respectively).
Assemblages A and C have a wide bathymetric range, from 198 to 1124 m and from 49 to 1100 m, respectively; on the contrary, group B has a narrow and relatively deep bathymetric range (733 to 866 m). The much larger bathymetric distribution of the assemblages A and C may partially explain their high biodiversity (11 and 25 species, respectively) compared to that of group B, which consists of only two species.
Interestingly, all three assemblages are characterized by species with a wide bathymetric range (cf. Table 1), therefore depth seems not to be the reason for the observed grouping. All species characterizing assemblage A are eurybathic (cf. , as it is the case for group C (except for A. operculata and S. liouvillei, which are absent from the shallowest part of the continental shelf, and R. antarcticum, which is restricted to the continental shelf). In the case of S. frondosus, the species that characterizes assemblage B, even when it has a wide bathymetric distribution, having been reported between 321 and 2283 m, it is clearly a species with deep-water affinities.
Assemblage C is mainly characterized by A. operculata, S. exochus, S. liouvillei, and O. terranovae. Except for S. exochus, which usually forms bushy colonies or grows epibiotic on other hydroids, the species form erect colonies. This group presents the highest hydroid diversity with 25 species, many of them epibiotic species that found a suitable substrate in two of the species present in this assemblage, O. terranovae and S. nana, which are usually found hosting other hydroid species (see above).
Assemblage A, the second in hydroid diversity, with 11 species, is characterized by four non-epibiotic species: B. subrufa, H. incertus, S. nana, and S. nonscripta. This group clearly differs from that of assemblage C by the absence of typically epibiotic species. The hydroid fauna found is this group is represented by species forming erect (e.g., H. incertus, S. nana, S. pachyclada) or bushy colonies (e.g., Staurotheca compressa, Staurotheca densa, S. nonscripta, S. glacialis).
Assemblage B is particularly interesting because it is formed almost exclusively by a single species, S. frondosus; there is another species, Tubularia sp.1, which was once found epibiotic on one colony of S. frondosus. The solitude of S. frondosus in this assemblage could be indicating an environment not very suitable for benthic hydroids other than this species. It forms large colonies up to 215 mm high, which have been reported epilithic on stones and epibiotic on coral, bryozoans, sponges, and mollusc shells (Peña Cantero 2017, 2019; present study). Peña Cantero (2021) found another benthic hydroid assemblage, in relatively shallow waters (from 36 to 133 m) of Point Géologie Archipelago (Adélie Land), clearly dominated by another species forming large colonies, O. terranovae. The structure of the hydroid assemblage in both cases is, however, completely different. As stated above, S. frondosus, despite being an abundant species forming large colonies that add spatial complexity to the community and could potentially be a habitat builder, is almost completely deprived of epibionts. Among the many colonies examined in the present study and in Peña , only once was another hydroid observed growing on it. Conversely, O. terranovae has been demonstrated to be a real habitat builder and hotspot of hydroid diversity (Peña Cantero 2021), with up to 34 different species of hydroids reported on it.
Only two previous studies have addressed the study of Antarctic benthic hydroid assemblages. Peña Cantero and Manjón-Cabeza (2014) investigated hydrozoan assemblages from shelf bottoms of Peter I (from 86 to 380 m) and bottoms from the deepest part of the continental shelf and the continental slope (from 426 to 2043 m) of the Bellingshausen Sea. They found three well-defined groups (deep-sea, continental slope, and shallowest stations), with depth apparently being the main underlying factor. Peña Cantero (2021) studied hydroid communities of Pointe Géologie Archipelago, from a narrow, relatively shallow, bathymetric range (from 19 to 172 m). He found four distinct assemblages, one of them clearly subdivided into two groups. In this case, depth seemed not to play a clear role in the assemblages found.
In the present study, of the three groups obtained, only assemblage B could be partially determined by depth, as the species characterizing it seems to be restricted to deep waters. Nevertheless, the virtually complete absence of other hydroid species in this assemblage clearly points to the influence of other factors too, because many Antarctic benthic hydroids are eurybathic and therefore depth would not be excluding them from this group. The absence of an appropriate substrate seems to be one of the main factors (evidenced by the absence of epibionts on S. frondosus).
In the case of Pointe Géologie Archipelago assemblages, substrate was clearly part of the underlying factors, thanks to the key role played by O. terranovae as habitat builder and hotspot of hydrozoan diversity (Peña Cantero 2021). The presence of this species, together with S. nana, in assemblage C might partially explain its hydroid diversity, which is characterized by a relatively rich fauna of epibiotic hydroids. Table 1 shows the bathymetric range of the species studied, as well as their entire known depth distribution. All the bathymetric groups proposed by Peña , but for the contingent of species restricted to approximately the shallowest 30 m (a depth range not sampled in the study), are represented in the collection. Of the five groups with representation, the clearly dominant one, with 19 species (51.3%), is that formed by eurybathic species, i.e., those distributed from the shallowest levels of the continental shelf to beyond the shelf-break. It is closely followed by the group composed of species extending from below the shallowest levels of the continental shelf to beyond the continental shelf-break (12 species, 32.4%). The other three groups are much less represented: R. antarcticum is distributed across the whole continental shelf; Eudendrium megaloarmatus sp. nov. and Schizotricha frigida sp. nov. are shelf species absent from the shallowest bottoms; and, finally, Nemertesia gelida sp. nov., Symplectoscyphus pseudofrondosus sp. nov. and Symplectoscyphus tortuosus sp. nov. belong to the group of species exclusively inhabiting bottoms beyond the continental self-break.

Geographical and bathymetric distribution
The Antarctic benthic hydroid fauna is characterized by a high level of endemism at the species level (Peña Cantero 2014b), which is also supported by the results of the present study. Twenty-six out of the 37 identified species are endemic to Antarctic waters (Table 1). These species belong to two categories: the dominant group, with 20 species (54.1%), is that formed by the species with a circum-Antarctic distribution (i.e., species reported around the Antarctic continent); the second group, with six species (16.2%), consists of species with East Antarctic distribution (i.e., restricted in their distribution to East Antarctic waters). The level of endemism in the present study (70.3%) is slightly higher than that found in other East Antarctic areas [e.g., 65% off Queen Mary Coast (Peña Cantero 2014b), 65% off Adélie Land (Peña Cantero 2021), or 68% in the Ross Sea and neighboring areas (Peña Cantero 2019)] and than the general endemism rate for Antarctic hydroids (65%, Ronowicz et al. 2019). However, it is lower than in other areas [e.g., 81% off Low Island, West Antarctica (Peña Cantero 2013)]. Apart from the endemic species, it is also remarkable that nine species are restricted in their distribution to Antarctic and sub-Antarcic waters: four Antarctic Patagonian, three Antarctic Kerguelen, and two Pan-Antarctic species. Thus, it is noteworthy that almost the entire fauna studied is restricted to Antarctic or Antartic/sub-Antarctic waters (35 species, 94.6%), as reported previously for other areas. Only two species (L. dumosa and Stegopoma plicatile) have a wider distribution, being also present outside these waters. However, these species are likely to constitute cryptic species complexes with hidden diversity, and their nearly cosmopolitan distribution being an artifact of incomplete taxonomic knowledge (e.g., Schuchert 2017).

Author contributions
The author carried out all the different aspects of the present study.
Funding Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.

Conflict of interest
The author declares that there is no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.