Introduction

Among the different levels of organization (from genes to ecosystems), the number of species observed in a given area is the most common measure of diversity (Gaston, 2000; Appeltans et al., 2012). Species diversity is closely connected with ecosystem functioning. With the increasing diversity, both species and species functions increase, and ecosystems become more complex over time (Boero & Bonsdorff, 2007). Any change affecting a biological system causes a progressing chain of transformations to retain its stability (Red Queen hypothesis, Van Valen, 1973). Currently, the need to explore marine biodiversity is heightened by the effects of global warming, which is more apparent in high latitudes than anywhere else on earth. Polar regions are warming rapidly (Walczowski & Piechura, 2006; Paolo et al., 2015), and the consequences are already clearly noticeable. Since 1950, the Arctic sea ice loss rate has reached 7.8%/decade (Stroeve et al., 2007), and the minimum sea ice extent ever observed in the Arctic was in September 2007 and 2008 (Wang & Overland, 2009). It is predicted that by 2037, the Arctic may be ice free in late summer (Wang & Overland, 2009). In some regions of the Southern Ocean, there is also evidence of climate change effects, e.g. the ice shelves in the Amundsen and Bellingshausen Seas have already lost up to 18% of their thickness in less than 20 years (Paolo et al., 2015).

Despite many shared attributes, such as the geographical position of the poles, low temperatures, ice coverage, and extreme seasonality in light regimes, these regions differ in many aspects (Table 1). The Southern Ocean is adjacent to all the great world oceans and surrounds the Antarctic continent, while the Arctic Ocean is a mediterranean basin, isolated to a great extent by land masses. The oceanic surface area of the Antarctic is more than twice that of the Arctic. The Arctic has broad and shallow continental shelves, which contrasts the deep and narrow shelves of the Antarctic. The glacial histories of both polar regions are strikingly different. The Antarctic region, a frigid environment, is much older than the Arctic. The Southern Ocean gained its cold-water attribute during the last 25–34 million years when the current regime of a cold ring ocean developed, with cold and stable conditions persisting for the last 10–17 million years (Clarke & Johnston, 1996). Conversely, the Arctic, a cold-water system, has a shorter and more dynamic evolutionary history. According to recent findings, approximately 17 million years ago, the Arctic Ocean became colder, and perennial ice cover developed approximately 13 million years ago (Krylov et al., 2008). However, modern Arctic biota are believed to be much younger and in a phase of colonization processes (Dunton, 1992; Piepenburg, 2005; Briggs, 2007) because, during the Last Glacial Maximum (approximately 18,000 years ago), a vast area of the Arctic continental shelf was covered by glaciers or exposed due to the lowered sea level (Nørgaard-Pedersen et al., 2003; Clarke & Crame, 2010), which must have resulted in the massive eradication of the marine shelf fauna. A comparison of diversity between the two polar regions, climatically similar but differing in geological history and shelf geomorphology, may reveal different and unique patterns of biodiversity.

Table 1 Comparison of oceanographical features of the polar oceans (after Dayton et al., 1994; Clarke & Harris, 2003; Eastman, 1997; Griffiths, 2010; Jakobsson, 2002; Krylov et al., 2008; Stonehouse, 1989; Starmans et al., 1999)
Full size table

The class Hydrozoa, particularly the two dominant orders, Leptothecata and Anthoathecata, which comprise 90% of all taxa, are here used as a model group. Hydrozoans are widely distributed, relatively rich in species and quite well represented in polar regions. The group is among the first explored in early polar diversity studies (Peña Cantero, 2014a; Ronowicz et al., 2015). They are important constituents in polar marine ecosystems, locally reaching high biomass and species richness (Włodarska-Kowalczuk et al., 2009; Clarke & Johnston, 2003). Pelagic hydrozoans are efficient predators, while benthic hydroids are successful suspension feeders that constitute an important link in bentho-pelagic coupling in the Arctic (Orejas et al., 2013) and the Antarctic (Gili et al., 1996). They promote diversity and enrich the structural complexity of the sea bottom by providing three-dimensional substrata (Gili & Coma, 1998; Ronowicz et al., 2013). By having two stages in their life cycle (pelagic medusa and benthic polypoid phases), hydrozoans offer the opportunity to study diversity aspects from two different perspectives: pelagic and benthic.

This study constitutes the first biodiversity comparison of the polar regions involving Hydrozoa and establishes a novel common framework that could be applied when dealing with other polar groups. Updated catalogues of the species inhabiting the Arctic and the Southern Ocean are generated, with their depth ranges, frequency of occurrence and information on missing or possibly extinct species in the region (not recorded for the last 40 years and 100 years). Knowledge of up-to-date, regional species catalogues is important for policy and management goals to evaluate priorities for conservation and sustainable use.

The main aim of this study is to examine whether and how the differing geological histories of the two polar regions, namely, the Arctic and the Antarctic, influence hydrozoan species/genus/family richness and endemism rate.

We present a comparative investigation based on the compilation of current knowledge on hydrozoan species composition and the distribution of two major orders (Leptothecata and Anthoathecata) from both polar regions. Such comprehensive species lists with distributional data allowed us to answer several research questions: (i) How well represented are hydrozoans in the polar regions compared to other marine invertebrates? (ii) Do differences in the continental shelves’ geomorphology affect the bathymetric distribution trends? (iii) Are the life history strategies of hydrozoans similar in both polar regions? (iv) Do the species with an effective long-living dispersal stage (i.e. medusa) have wider distributions in both regions? (v) Are there true bipolar taxa? (vi) Are there any possibly extinct species in the polar regions?

By answering the above questions, we will be able to provide solid baseline information about hydrozoan species richness and composition, which will act as a valuable tool for the future evaluation of climate change effects in polar regions. This study will also contribute to understanding the processes shaping biodiversity, namely, geological history or sea bottom geomorphology.

Materials and methods

Study area

The Arctic Ocean covers an area of 14 million km2 and is the shallowest of all the oceans. Its mean depth is 1201 m. Continental shelves cover the largest sector, which constitute 53% of the total Arctic Ocean area (Jakobsson, 2002). It is a nearly landlocked oceanic entity consisting of the deep oceanic central Arctic Basin; the broad continental shelves of the Barents, Kara, Laptev, East Siberian, Chukchi, and Beaufort Seas; the White Sea; and the narrow continental shelf of the Canadian Arctic Archipelago and northern Greenland (Jakobsson, 2002). It is connected to the Pacific Ocean through the shallow Bering Strait and to the Atlantic Ocean through the deep Fram Strait, the Barents Sea and the Canadian Archipelago. The limit of the Arctic region is not clearly defined. For the purpose of this study, we considered the boundary to be that used by the Arctic Monitoring and Assessment Program (AMAP, 1998) (with the Bering Sea, southern Greenland and Iceland included) (Fig. 1). The sub-Arctic region includes the temperate waters of the North Atlantic and the North Pacific that are connected to the Arctic ecoregion (sensu Spalding et al., 2007) and encompasses the Faroe Plateau, western Norway, the Gulf of St. Lawrence, south Newfoundland, the Aleutian Islands and the Gulf of Alaska.

Fig. 1
figure 1

Map of the Arctic and Antarctic regions showing polar areas demarcated with dotted lines

Full size image

The Southern Ocean’s area is approximately 34.8 million km2. The shelves around Antarctica are narrow and relatively deep, with an average of 450–500 m deep and extreme depths of 1000 m in some regions (Clarke & Johnston, 2003; Gili et al., 2016). The limits of the Antarctic region are better defined than the Arctic region (at least in the pelagic system): the existence of the Polar Front (Antarctic Convergence Zone), a major encounter of water masses, has been considered “one of the strongest natural boundaries in the world ocean” and an effective barrier (filter) for dispersal (Crame, 1999). Thus, the area south of the Polar Front (including the whole Scotia Arc and Bouvet Island) is considered here, which is in agreement with Hedgpeth (1969) and Dell (1972) as well as recent contributions on the biogeography of Antarctic hydrozoans (Soto Àngel & Peña Cantero, 2017; Mercado Casares et al., 2017). The northern limits of the sub-Antarctic Patagonian region established by Briggs (1974) and later used by Peña Cantero & García Carrascosa (1999) have been followed: 35°S latitude in the eastern sector and 42°S in the western sector. Marion Island, Macquarie Island, Îles Crozet or Îles Kérguelen are considered sub-Antarctic (after Clarke, 1996).

Data compilation and analyses

To compile the database of polar hydrozoans, the datasets published by Ronowicz et al. (2015) for Arctic Hydrozoa and those by Soto Àngel & Peña Cantero (2017) and Mercado Casares et al. (2017) for Antarctic Hydrozoa were used. These datasets were based on the compilation of all records existing in the scientific literature from each polar region. All the species belonging to the hydrozoan orders Leptothecata and Anthoathecata were extracted. The information was corrected and updated with the most recent publications, and rare taxa that were neglected in previous catalogues were added. The following accompanying data were also incorporated for each species inventoried: life history strategy, biogeographical pattern of distribution frequency of occurrence, bathymetric range, and if it was sampled within the last 40 years. The databases for Arctic and Antarctic hydrozoan species are incorporated as supplementary tables (Appendices 1a, 2a). Species with uncertain taxonomic status (species inquirenda and doubtful species) were excluded from subsequent analyses, but the corresponding information is provided in each database. In addition, the comprehensive databases with the bibliography on polar hydrozoan research and local synonymies are included as supplementary tables (Appendices 1b, 2b).

To assess the diversity contribution of polar regions to the total number of species described, current data on the species numbers of the main taxonomic groups were extracted from the Register of Antarctic Marine Species (RAMS, De Broyer et al., 2018) and the Arctic Register of Marine Species (ARMS, Sirenko et al., 2018). Global information at the species level was extracted from the World Register of Marine Species (WoRMS, Editorial Board, 2018) and its sub-registers (Global Species Databases, GSD) when available. For Hydrozoa, the analysis was also undertaken at the family and genus levels. Unaccepted, extinct or exclusively freshwater taxa were removed from every inventory.

For the study of missing species, the last record of each species in each polar region was tracked in the literature. The date when samples were obtained (not the year of publication) was used. Species not collected (so-called missing species) for the last 40 years (since 1977) and 100 years (since 1917) were extracted from the database for further analysis. Species recorded within the period analysed that could not be determined with certainty were not considered in the analysis.

The zoogeographical affinity of both Arctic and Antarctic species was simplified to enable the comparison between the two regions. Three categories were considered: species occurring exclusively either in Arctic or Antarctic waters (E—endemic); species also known from sub-Arctic or sub-Antarctic waters (SA—distributed from the sub-polar to the polar region); and species also present outside each polar–sub-polar region (W—widely distributed species).

In terms of life cycle strategies, species were separated into two major groups to mirror their dispersal potential (Gibbons et al., 2010): benthic (B) and meroplanktonic (M) species. The former group (B) corresponds to the medusa-lacking species, with the exception of a single representative with benthic hydromedusae [i.e. Staurocladia charcoti (Bedot, 1908)], and the latter (M) includes the species producing free-swimming medusae. These groups enabled the comparison of the dispersal capabilities of the two strategies, with the working hypothesis that species with a medusa stage in their life cycle (M) are more widely distributed than the species without (B), and, consequently, endemism in M would be correspondingly lower than in B.

Species were ranked according to their frequency of occurrence based on the historical records of each region. The unit of measure used was the number of stations, which refers to a unique combination of latitude, longitude, depth and time; therefore, a single scientific contribution can encompass several stations/records. The categories used include (1) rare species, 1–9 records; (2) rather rare species, 10–19 records; (3) rather common species, 20–29 records; (4) common species, 30–69 records; and (5) very common species, over 70 records. For the Arctic species for which the number of records could not be estimated, we adopted Naumov’s (1969) suggestions.

To compare the bathymetric distribution of the Arctic and Antarctic hydrozoans, a two-way approach was employed. On the one hand, the bathymetric patterns proposed by Peña Cantero (2004) were used. The groups were unevenly distributed in depth to reflect the different geomorphologies of the Antarctic and Arctic shelves: the shelf slope begins at depths from approximately 450–500 m in the Antarctic (Dayton, 1990; Griffiths, 2010), while in the Arctic, the continental shelf break is generally defined by 200-m isobaths (Johnson, 1990). The bathymetric patterns used include (1) shallow (S)—depth less than 30 m; (2) continental shelf (CS)—depth between 30 m and the shelf break; (3) shallow and continental shelf (S + CS)—depth shallower than 30 m to the shelf break; (4) continental shelf to deep sea (CS + DS)—depth from 30 to beyond the shelf break; (5) deep sea (DS)—depth beyond the shelf break; and (6) shallow, continental shelf and deep sea (S + CS + DS)—depth shallower than 30 m to beyond the shelf break. Six bathymetric ranges were defined, and a presence–absence data matrix was created with the species inventoried for each range: 0–30 m, 30–200 m, 200–500 m, 500–1000 m, 1000–3000 m, and deeper than 3000 m. The Sørensen similarity index was subsequently calculated, and a cluster analysis (hierarchical agglomerative linkage by group average) was performed to compare the relationship among bathymetric ranges between regions by means of PRIMER v.7 (Clarke & Gorley, 2015). Species with unknown bathymetric distributions were not included in the analysis.

For all analyses, a Pearson’s Chi-square test was used to assess differences in the occurrence of hydrozoan species in the different categories and to measure whether the proportions of the groups differed between the two polar regions. Yates’s correction was applied when applicable.

Results

Diversity

We compiled a total of 8653 records (4192 from the Antarctic and 4461 from the Arctic) of polar Hydroidolina of the orders Leptothecata and Anthoathecata (henceforth, collectively named hydrozoans) from 254 references published since 1875 (until February 2018). The complete inventory includes 469 species belonging to 136 genera and 54 families (excluding doubtful species and species inquirenda, see Appendices 1a and 2a). Therefore, polar regions harbour 14% of the known species of Leptothecata and Anthoathecata (Table 2). At a higher taxonomic level, the number of represented taxa increases; one third of the genera, and more than half of the families known globally have representatives inhabiting polar waters (Table 2). At the species level, the Antarctic region has a higher proportion of Leptothecata to Anthoathecata, and at both polar regions, this ratio is higher when compared to global data. However, at the genus and family levels, the proportion varies: while the diversity at those taxonomic ranks is higher in the Arctic Leptothecata (ratio > 1), the Antarctic and global data show opposite patterns (ratio < 1) (Table 2).

Table 2 Diversity of hydrozoan fauna (Leptothecata and Anthoathecata) in the Arctic, Antarctic, polar waters (Arctic + Antarctic) and worldwide (extracted from WoRMS)
Full size table

A total of 233 species of hydrozoans (Leptothecata and Anthoathecata) belonging to 95 genera and 41 families are recorded for the Arctic region (see Table 2 and Appendix 1a). The mean number of species per family is 5.7, with Sertulariidae as the most speciose family, which is widely represented by 50 species (21%) in seven genera. Other speciose families are Haleciidae, with 21 species (9%) belonging to a single genus, and Campanulariidae, with 15 species (6%) in eight genera. The mean number of species per genus is 2.5, with the largest number for Halecium (21 species) and Thuiaria (20 species).

The Antarctic inventory comprises 252 species belonging to 88 genera and 45 families (see Table 2 and Appendix 2a). The mean number of species per family is 5.6, with Symplectoscyphidae as the most speciose, with 34 species (14%) in two genera. Other speciose families are Kirchenpaueriidae, with 26 species (10%) in one genus; Staurothecidae, with 23 species (9%) in one genus; Stylasteridae, with 21 species (8%) in 10 genera; and Haleciidae and Schizotrichidae, each with 14 species (6%) belonging to a single genus. The mean number of species per genus is 2.9, with the largest number for Oswaldella (26 species), Symplectoscyphus (24 species) and Staurotheca (23 species).

There are some remarkable similarities as well as differences between the Arctic and Antarctic hydrozoan fauna, which are summarized in Table 2. Species richness and diversity at the family level are higher in the Antarctic region but lower at the genus level. The distribution of species between the two orders is similar: Leptothecata predominates, representing 65% of the species in the Arctic and 69% in the Antarctic. Six families (between 6 and 7% of the total) comprise more than 50% of the species in both regions, although the families differ, with the exception of Haleciidae (Fig. 2). Many families in polar regions are represented by a single genus: 20 (49%) in the Arctic and 29 (64%) in the Antarctic, with a mean number of genera per family of 2.3 and 2.0, respectively, which is lower than the global pattern (Table 2). Conversely, the mean number of species per family in both polar regions is considerably lower than for worldwide hydrozoans. In the Antarctic region, 18 families (40%) are represented by a single species, while in the Arctic, this is the case for 11 families (28%). Similarly, most genera from polar regions are represented by only one species: 52 (55%) from the Arctic and 53 (60%) from the Antarctic, with a mean of 2.5 and 2.9 species per genus, respectively, which is considerably lower than the global mean of 8.1. Halecium is the only speciose genus (> 10 species) shared between both regions. Thirty-two families (60%) are common in both regions. Among them, Campanulariidae, Haleciidae and Lafoeidae are the most speciose taxa (Fig. 2). At the genus level, 47 (34%) taxa are shared between the Arctic and Antarctic regions, while only 16 species (3%) are known to occur in both polar oceans. Among the shared species, most are widely distributed, and only a few are reported exclusively from polar regions (see below for a "Discussion" on bipolarity).

Fig. 2
figure 2

Number of species and genera per family in the Arctic and the Antarctic. Taxa with an uncertain systematic position and families represented by a single species in each region are not shown

Full size image

Distribution and life cycle strategy

The zoogeographical affinity of the hydrozoan fauna differs considerably between the Arctic and Antarctic (Pearson Chi-square test: χ2 = 107.56, df = 2, P < 0.05). In addition, there are considerable differences between the distribution of benthic and meroplanktonic taxa in both regions. Endemism is much more pronounced in the Antarctic than in the Arctic, with 144 (58%) species restricted to the Antarctic region, while 47 (20%) are known exclusively from the Arctic waters (Fig. 3). Sixty-six percent of the species have more widespread distributions in the Arctic, extending outside Arctic and sub-Arctic waters, while only 20% of taxa occur outside sub-polar and polar waters in the Antarctic. Antarctic benthic representatives show the highest endemism (Fig. 3), with 133 (65%) species of benthic hydroids inhabiting exclusively the Southern Ocean. Meroplanktonic taxa show a wider distribution than benthic forms in both regions, but endemism in Antarctic representatives is higher than in the Arctic.

Fig. 3
figure 3

Number and percentage of species according to life cycle strategy and distributional patterns. Endemic: exclusively polar species known either from the Arctic or the Antarctic; SA polar + sub-polar species known from either the sub-Arctic–Arctic or the sub-Antarctic–Antarctic, W widely distributed, species also occurring outside polar–sub-polar waters

Full size image

At a higher taxonomical level, endemism is extremely low in both polar regions. To date, there are no endemic hydrozoan families known from polar waters. Among the 95 genera present in the Arctic Ocean, only the recently discovered monotypic genus Sympagohydra may be considered endemic for the Arctic region. Similarly, the monotypic Mixoscyphus is the only true endemic genus from the Southern Ocean. In the Antarctic, there are a few more genera with a high level of species endemism; for example, of the 10 known species of Antarctoscyphus and the 27 known species of Oswaldella, only two and one species, respectively, have been found outside the Southern Ocean in sub-Antarctic regions.

The common pattern of the predominance of benthic taxa over species with medusa stage is consistent for each zoogeographical group in both the Arctic and the Antarctic, with the exception of widely distributed Antarctic taxa that have more meroplanktonic representatives (Fig. 3). The majority of the hydrozoan fauna in both polar regions are deprived of a medusa stage and spend their entire life cycle as benthic forms, with 175 (75%) species in the Arctic and 204 (81%) in the Antarctic. In addition, a single benthic hydromedusa belonging to Anthoathecata is known from the Antarctic (i.e. Staurocladia charcoti). Accordingly, the presence of a medusa stage occurs in 58 (25%) of the Arctic hydrozoan species and 48 (19%) of the Antarctic species.

Frequency of occurrence

In the investigated regions, a great majority of the species surveyed have been reported rarely, with less than 10 records. A higher percentage of rare species is observed in the Antarctic for both life cycle strategies (i.e. benthic and meroplanktonic) (Fig. 4), constituting 63% of the species; in the Arctic, they represent 49%. The distribution in different classes of frequency of occurrence varied significantly between the Arctic and the Antarctic (Pearson Chi-square test: χ2 = 13.0817, df = 4, P < 0.5). The polar regions differ in the frequency of occurrence of life cycle strategies: for Antarctic hydrozoans, the rarity of meroplanktonic species is higher than the rarity of benthic species (the highest among all the groups considered here), whereas the opposite pattern is observed in the Arctic hydrozoans. The contingent of common and very common benthic species is identical for both regions (18%), which is larger than the contingent of meroplanktonic species, particularly in the Arctic. According to the results, very few species from both regions are regarded as very common. Among them, the most frequent (> 100 records) Antarctic hydrozoans are the hydromedusa Calycopsis borchgrevinki (Browne, 1910) and the hydroids Antarctoscyphus elongatus (Jäderholm, 1904), Antarctoscyphus spiralis (Hickson & Gravely, 1907), Billardia subrufa (Jäderholm, 1904) and Symplectoscyphus glacialis (Jäderholm, 1904). Arctic species with the highest frequency of occurrence are the hydroids Grammaria abietina (Sars, 1850), Halecium muricatum (Ellis & Solander, 1786), Lafoea dumosa (Fleming, 1820) and Symplectoscyphus tricuspidatus (Alder, 1856).

Fig. 4
figure 4

Number and percentage of species according to frequency of occurrence and life cycle strategy

Full size image

Missing species and recent discoveries

The last record of each species was traced in the literature. There have been no records of 89 species (74 benthic and 15 meroplanktonic, 38% in total) in the Arctic and 64 species (49 benthic and 15 meroplanktonic, 25% in total) in the Antarctic in the last 40 years (Fig. 5). The differences in the contribution of present and absent species between the regions are significant (Pearson Chi-square test: Yates’s χ2 = 8.63, df = 1, P < 0.05). Several polar representatives are extremely rare and have not been noted in the polar regions for the last 100 years. Arctic representatives include 16 species: Corymorpha carnea (Clark, 1876), Halecium harrimani Nutting, 1901, and Halecium leave Kramp, 1932 (all three known only from their type locality), Eudendrium caricum Jäderholm, 1908, Kirchenpaueria fragilis (Hamann, 1882), Plumalecium plumularioides (Clark, 1877), Lafoea symmetrica Bonnevie, 1899, Thuiaria pinaster (Lepechin, 1783) and Stylaster norvegicus (Gunnerus, 1768), among others. Likewise, 16 species have not been reported from the Antarctic region for the last 100 years. This is the case for Bythotiara drygalskii Vanhöffen, 1912, Eudendrium cyathiferum Jäderholm, 1904, Eudendrium jaederholmi Puce, Cerrano & Bavestrello, 2002, Halecium pallens Jäderholm, 1904, Sertularella nuttingi Billard, 1914, Tubularia antarctica Hartlaub, 1905, Tubularia hodgsoni Hickson & Gravely, 1907 and Rhysia halecii, among others (all known exclusively from their type localities).

Fig. 5
figure 5

Number and percentage of species present and absent during the last 40 years according to life cycle strategy

Full size image

In recent years, (since 2000), five new species of benthic hydroids have been described from the Arctic region, which corresponds to 2% of the total known species richness. In the Antarctic, this number is considerably higher, with a total of 46 new hydroids and one new hydromedusa, representing 18% of the known representatives.

Bathymetry

The vertical distribution of the hydrozoan species (Fig. 6) differs significantly between the Arctic and Antarctic regions (Pearson Chi-square test: χ2 = 91.53, df = 5, P < 0.01). Among the ranges considered here, the highest species richness is found between 30–200 m in the Arctic and 200–500 m in the Antarctic, each with an almost equal number of species (Fig. 6). Below 200 m depth, there is an increase in species richness by a factor of 1.2 in the Antarctic, but a decrease by a factor of 1.4 in the Arctic. Beyond 500 m, the species richness drops by a similar factor in both regions (1.7 in the Arctic and 1.6 in the Antarctic). The difference increases below 1000 m, with a reduction factor of 1.6 in the Arctic and 2.5 in the Antarctic. Below 3000 m depth, very few species are known in both regions, but slightly more are known in the Arctic (Fig. 6).

Fig. 6
figure 6

Number of species from each polar region according to depth ranges

Full size image

The continental shelves (including the shallowest region) are the richest sectors, with nearly identical species counts in the Arctic (202 species, 91%) and the Antarctic (219 species, 89%) (Fig. 7). In contrast, the number of species whose occurrence is restricted to the shelf (“S”, “CS” and “S + CS”) are unevenly distributed, with 79 (35%) species from the Arctic and 122 (50%) from the Antarctic waters (Fig. 7). The shallowest depth range (0–30 m) exhibits contrasting patterns between the study areas: shallow waters from the Arctic are very rich, with 70% of the total species present and with more than twice the species richness of the Antarctic waters (Fig. 6). On the other hand, only 30% of the Antarctic species have been reported from this narrow depth range (Fig. 6). Few species from each region are restricted only to the shallowest range: 14 (19%) from the Antarctic waters and 17 (11%) from the Arctic. The deep waters (regions extending beyond the shelf break) are slightly richer in the Arctic, with 144 (65%) species, compared with 124 (50%) species from the Southern Ocean. The number of strictly deep-water taxa is similar among regions, with 21 (9%) Arctic and 27 (11%) Antarctic species (Fig. 7).

Fig. 7
figure 7

Percentage and number of species according to bathymetric categories. S shallow: swallower than 30 m, CS continental shelf: 30–200 m in the Arctic, 30–500 m in the Antarctic, S + CS shallow and continental shelf: swallower than 30 m to swallower than shelf break, CS + DS continental shelf and deep sea: from continental shelf (deeper 30 m) to beyond the shelf break, DS deep sea: beyond the shelf break; eurybathic (S + CS + DS) depth shallower than 30 m to beyond the shelf break

Full size image

Regarding the bathymetric patterns of species richness (Fig. 7), the continental shelf excluding shallow waters (CS) is the most speciose zone in the Antarctic region, with 84 (34%) species. In the Arctic, however, hydrozoans are poorly represented in this zone, with only 13 (6%) species. In contrast, the category of shallow waters, continental shelf and beyond shelf break (S + CS + DS) is the most species-rich in the Arctic, with 91 (41%) records, while 37 (15%) are known from the Antarctic (Fig. 7).

Among all the Antarctic taxa surveyed, 138 species have a depth range over 300 m, 99 species occur over a 500 m depth range, and 47 species occur over 1000 m. In Arctic waters, this is the case of 112, 81 and 46 species, respectively. There is only one species from the Southern Ocean with a depth range over 3000 m, which is absent from the shallowest waters. In contrast, six representatives from Arctic waters occur in this wide bathymetric range, including the shallowest zones.

The hierarchical analyses of the bathymetric ranges surveyed for each region show different clustering (Fig. 8). The most distinctive bathymetric group in both polar regions inhabits waters below 3000 m depth, with very low similarity with the remaining groups. Shallow water hydrozoan assemblages are particularly distinctive in Antarctic waters. On the contrast, in the Arctic, the shallowest category shows a high similarity (> 80%) with the subsequent bathymetric group (i.e. 30–200 m). The highest similarity (approximately 80%) in Arctic waters is found between 0–30 and 30–200 m, while in the Antarctic, the highest similarity occurs between 30–200 and 200–500 m. Both groups coincide with the continental shelf extension within each region. A transition is further observed from both clusters (a similarity of 60–70%), between 0–200 and 200–500 m in the Arctic and 30–500 and 500–1000 in the Antarctic, both encompassing their corresponding shelf break. The deep hydrozoan fauna inhabiting 500–1000 m depths shows a higher similarity with the deeper groups (i.e. 1000–3000 m) in the Arctic but with shallower groups in the Antarctic.

Fig. 8
figure 8

Cluster analyses of bathymetric ranges based on the Sørensen similarity index for the Arctic (above) and the Antarctic (below) data

Full size image

Discussion

As our analysis of polar hydrozoans indicates, the differing geological histories of the Arctic and the Antarctic (see “Introduction”) influence species/genus/family richness as well as endemism rates within this group of invertebrates. These two polar regions have a number of similarities with regard to environmental conditions, including low water temperatures, the presence of ice or very distinct seasonality due to the existence of polar day and night. However, despite these similarities, the species/genus/family composition differed between regions, and the endemism level was higher, as expected, in the older system (Antarctica).

Arctic–Antarctic diversity comparison

The common assumptions regarding diversity and endemism in polar regions include a higher faunal species richness in Antarctic marine waters than in Arctic waters (Jażdżewski et al., 1995; Schäfer, 2007; Sirenko, 2009; Pabis et al., 2015 and literature cited). Similarly, high endemism is a genuine biogeographical and evolutionary feature of the Antarctic benthos, ranging from 45 to 75% depending on the group (cf. Griffiths, 2010), whereas the Arctic Ocean exhibits much lower levels (ca. 20%, Briggs, 2007). The results obtained in the present study using hydrozoans as a model group are congruent with some of these ideas.

The present study shows a slightly higher hydrozoan diversity in the Antarctic (Table 2). This is particularly apparent when comparing the number of hydrozoans with a benthic life history strategy. In other marine invertebrates, the differences in species richness between both regions were more evident (Knox & Lowry, 1977; Piepenburg et al., 1997; Schäfer, 2007). The current information, compiled from the corresponding polar registers of marine species (De Broyer et al., 2018; Sirenko et al., 2018), shows that the number of known species in the Southern Ocean exceeds that in the Arctic for more than 3000 species (Table 3). A great majority of the surveyed taxa are considerably more speciose in the Antarctic, particularly acarines, brachiopods, cephalopods, hexactinellid sponges, isopods, pycnogonids, tanaids, and all classes within Tunicata and Echinodermata (Arctic/Antarctic species ratio < 0.4, Table 3). In contrast, a few groups are more diverse in the Arctic (e.g. Bivalvia, Caudofoveata, Decapoda, Nematoda and Oligochaeta).

Table 3 Comparison of polar diversities between the Arctic and the Antarctic
Full size table

The differences in the geological age and the degree of isolation of the two systems have been widely used to explain the higher number of species found in the Antarctic. Recent findings, however, have revealed that the modern oceanic circulation in the Arctic might date back 17 million years, with perennial sea–ice cover formed approximately 13 million years ago (Krylov et al., 2008), which is much earlier than previously thought (ca. 4 million years ago, Dayton, 1990; Piepenburg, 2005). Regardless, the current Arctic benthic fauna probably populated the shelf regions approximately 13,000 years ago, only after the last glacial maximum (Clarke & Crame, 2010). Before that time, the variation of the sea level by more than 100 m during the Pleistocene period resulted in either emergence or glacial cover of the shallow Arctic shelf, with total eradication of the shelf fauna during the glacial maxima (Clarke & Crame, 2010). In contrast, the corresponding elapsed time for Antarctic biota is approximately 25–34 million years, which is caused by the early separation of the continent from the Gondwana land mass and the establishment of the Antarctic Circumpolar Current, with cold and stable conditions persisting for the last 10–17 million years (Clarke & Johnston, 1996).

Another likely reason for the impoverished benthic habitats in the Arctic might be the homogeneity of soft bottoms due to elevated sedimentation inflow (Dayton, 1990); although rich in infaunal species compared to the Antarctic (Włodarska-Kowalczuk et al., 2007), soft bottoms are poor in epifauna (cf. Sirenko, 2009). This general pattern is supported by our analysis (Table 3): particularly speciose groups in the Antarctic belong to the epifauna, while those richer in the Arctic are mainly infaunal groups, which agrees with previous hypotheses (Knox & Lowry, 1977, Jażdżewski et al., 1995 and Sirenko, 2009). Abundant gravel and stones constitute an excellent substrate for the extraordinarily rich and conspicuous benthic suspension feeder assemblages in the Southern Ocean, which in turn increases habitat complexity and heterogeneity (Jażdżewski et al., 1995; Gutt & Piepenburg, 2003; Sirenko, 2009). These so-called animal forests persist in the Southern Ocean because of the constant nutrient supply due to resuspension processes (Gili et al., 1996; Orejas et al., 2000) and in turn serve as stable habitats for a wide array of vagile fauna (Starmans & Gutt, 2002; Gutt et al., 2016). This fact, coupled with low predatory pressure in the Antarctic communities (Aronson & Blake, 2001), has been regarded as a basis for the asymmetries in species richness observed between both regions (Pabis et al., 2015).

Additional drivers that have been identified as possible causes for the differences in species richness include the recent interchange with adjacent oceans, the age and variability of sea ice, the asymmetric impact from terrestrial freshwater input, and the stability of the Antarctic environment (Dayton et al., 1994). Nevertheless, some authors have regarded the assumption of greater species richness in Antarctic waters to be an overgeneralization (Piepenburg, 2005; Clarke & Crame, 2010). Others have questioned the validity of the comparisons when the species pools are independent, with scarce overlap in the species composition between regions, especially when the taxon analysed is unevenly represented (Pabis et al. 2015 and references therein). Moreover, some studies highlighted that the real diversity from both the Arctic (Piepenburg et al., 2011) and the Antarctic (De Broyer et al., 2011) are still underestimated and that the expected species richness in several groups might be comparable (Jażdżewski et al., 1995; present study). Remarkably, some authors have noted that the contrasting results obtained depended on the scale of the analyses, with no significant differences evident at the habitat scale but with higher species richness evident at larger scales (fjord, sea, or the whole region) in the Antarctic sectors than in the Arctic (cf. Starmans & Gutt, 2002; Pabis et al., 2015).

At the scale of a single sea, Sirenko (2009) also reported higher macrozoobenthic species diversity in the Antarctic Weddell Sea than in the Arctic Laptev Sea, emphasizing the low degree of study of the former. A similar trend between the Weddell and Greenland Seas was found for echinoderms (Piepenburg et al., 1997). At the scale of a fjord, analogous patterns have been described for amphipods, polychaetes, gastropods and isopods (Jażdżewski et al., 1995), polychaetes (Pabis et al., 2015), and tanaids (Błażewicz-Paszkowycz & Sekulska-Nalewajko, 2004). In contrast, Eastman (1997) reported more fish species in the Arctic, and Włodarska-Kowalczuk et al. (2007) found similar soft bottom polychaete diversity at the fjord scale. In addition, a comparison of planktonic communities did not show marked differences in species richness at the fjord scale (Walkusz et al., 2004). In spite of the aforementioned results and the fact that species richness might be greatly miscalculated depending on the research effort on the taxon analysed (De Broyer et al., 2011; Griffiths, 2010; Griffiths et al., 2011; Piepenburg et al., 2011), some interesting trends are derived from the present study.

Thus, the results are consistent with the general view of the asymmetric configuration between polar benthic assemblages, although the general pattern should not be extrapolated to all polar habitats or to the whole tree of life.

Contribution of polar fauna to global marine biodiversity

The percentage of polar species of each taxon with respect to the total species known globally (polar representativeness, henceforth known as PR) greatly differs between the taxa and regions analysed (Table 3). Remarkably, a few groups are very well represented in both polar regions (PR > ca. 10%), i.e. Amphipoda, Chaetognatha, Euphausiacea and Stenolaemata. Others are very well represented in the Southern Ocean (PR > ca. 20%) but barely known in the Arctic, i.e. Solenogastres, Pycnogonida, Appendicularia, Brachiopoda and Asteroidea. The contrary is true only for Caudofoveata (Table 3). Focusing on the most speciose taxa, some similarities between regions are observed. Polar regions share Amphipoda as the group with the highest number of species, while they are only the sixth most species-rich group worldwide. Conversely, other species-rich groups, such as Anthozoa, Decapoda and Gastropoda, are underrepresented in polar regions, which is likely due to physiological constraints, such as the difficulty of depositing CaCO3 in cold waters (Aronson et al., 2007). Polar waters harbour approximately 7% of the hitherto known marine invertebrate species, of which approximately 3% inhabit the Arctic and approximately 4% inhabit the Antarctic.

The class Hydrozoa (the compilation from the present study and including Siphonophora and Trachylina) shows high PR, which is above that of most of the taxa surveyed (Table 3). Within Cnidaria, Hydrozoa has the second highest PR value (just behind Staurozoa, a group with very few polar species) and the highest number of species (far more than the anthozoans) in the Arctic. In comparison with the remaining taxa, Arctic Hydrozoa are one of the groups with the highest PR value (Table 3) and the fifth in number of species. Regarding Antarctic representatives, Hydrozoa is the first taxon in PR value and number of species within the Cnidaria. It is also a taxon with high PR in the Antarctic, but this trend is less noticeable than in the Arctic. In terms of the number of species, hydrozoans are among the top ten invertebrate groups, even surpassing the so-called speciose Antarctic habitat formers (i.e. anthozoans, bryozoans, tunicates and sponges). Nevertheless, they are often underrated in both benthic and planktonic studies at the community level (e.g. Gutt & Piepenburg, 2003; Włodarska-Kowalczuk et al., 2009; Steinberg et al., 2015). The great plasticity of the life cycle and ecological strategies (cf. Cartwright & Nawrocki, 2010), including several holoplanktonic representatives, might be the reason behind this great representation.

Endemism and radiation

Hydrozoan species diversity from Antarctic waters is concentrated in a few speciose families with no endemism at the family level. They include a few genera each and are scarcely represented or even absent from other waters, including the Arctic (Fig. 2). In contrast, the most speciose families in the Arctic occur globally, including the Antarctic. Remarkably, similar results were observed in amphipods (Jażdżewski et al., 1995), echinoderms (Piepenburg et al., 1997), fishes (Eastman, 1997) and polychaetes (Pabis et al., 2015). At the species level, the endemism in Arctic hydrozoans (20%) is the same as that in the Mediterranean (Gravili et al., 2013). However, a level of endemism as high as the Antarctic (58%) has not been documented for Hydrozoa elsewhere. The disparity in endemism levels between both regions has been recorded in all previous works, yet the underlying processes are not completely clear. The Southern Ocean is effectively isolated from the surrounding waters by oceanographical barriers, the Antarctic Circumpolar Current and the Polar Front (Hassold et al., 2009). Conversely, the less isolated condition of the Arctic, which waters are constantly subjected to the inflow of warm waters from the boreal region and the subsequent dominance of boreo-Arctic species (George, 1977) has been regarded as the main reason for the low endemism values (Starmans & Gutt, 2002; Sirenko, 2009). This is also manifest in the polar planktonic communities, with most taxa showing a circum-Antarctic distribution in the Southern Ocean, while in the Arctic, there is a predominance of boreo-Arctic representatives (Walkusz et al., 2004; Weydmann et al., 2014; Mańko et al., 2015). According to our results, hydrozoans also show these patterns: the majority of the Arctic species are widely distributed, whereas Antarctic species are mainly restricted to the Southern Ocean (Fig. 3) and are consequently endemics. In addition, the opportunity for speciation has persisted longer in the ancient and stable Antarctic environment than in the young and unstable Arctic environment. In this regard, the Antarctic benthic assemblages mirror reminiscent features of the Paleozoic communities, with several relict taxa and other fauna that have evolved in situ (Clarke & Crame, 1989; Aronson & Blake, 2001; Gili et al., 2016).

Some authors have hypothesized that the elapsed time for speciation in the polar regions is responsible for regional variation in rates of endemism, with ‘stable’ deep sea areas harbouring more endemics than shelf and coastal marine areas, which have been free from bottom covering glacial ice for a relatively short amount of time (Payer et al., 2013 and literature cited). Our results do not confirm this hypothesis, as polar hydrozoan endemic species are found throughout the entire depth ranges in both the Arctic and the Antarctic (Fig. S1).

A higher species/genus ratio and a lower number of genera are observed in Antarctic waters than in Arctic waters (Table 2). Likewise, Schäfer (2007) reported similar patterns for polar cheilostome bryozoans and hypothesized that the results reflect the high degree of endemism in the Southern Ocean and the role of the Arctic Ocean in connecting the Pacific and Atlantic northern basins since the Miocene. The species/genus ratio in both polar regions is relatively low and appears to be a characteristic feature of polar hydrozoans. Indeed, most of the genera from both regions contain a single species (Appendices 1a and 2a). These features have also been observed for Antarctic bryozoans (Figuerola et al., 2017) and bivalves above 40° latitude (Krug et al., 2008). The putative rationalizations previously provided are not mutually exclusive and embrace the presence of genera containing species with low potential for speciation and/or as the result of strong intrageneric competition, which might limit the congeneric coexistence of species with similar ecological requirements (Gotelli & Colwell, 2001; Webb et al., 2002). Even so, a few genera are particularly prolific in polar waters. In the Arctic, Halecium and Thuiaria include 21 and 20 species, respectively, the latter corresponding to 40% of the valid species known worldwide (cf. Schuchert et al., 2018), of which 40% are in turn endemic from the Arctic. In Antarctic waters, this is the case for Oswaldella, Staurotheca, and Symplectoscyphus, with 26, 23 and 24 species, respectively. The representation of the first two genera in the Antarctic is very high, and the polar species constitute approximately 90% of the total species known worldwide, all with levels of endemism above 75%. In the Southern Ocean, the high radiation observed in a few nearly indigenous genera agrees with the high-latitude diversification hypothesis (Weir & Schluter, 2007), according to which just a few probably old genera are able to survive at high latitudes, showing adaptive radiation at higher rates than in the lower latitudes (Eastman, 1997). When and where this diversification took place and the origin of the pioneer taxa in the Southern Ocean are still unresolved issues. In this regard, recent publications suggest that the different hypotheses proposed by Knox & Lowry (1977) (i.e. evolution in situ, derivation from adjacent deep sea basins or dispersal from/to adjacent waters) might be co-occurring in Antarctic Hydrozoa (Miranda et al., 2013; Mercado Casares & Peña Cantero, 2018), since the different taxa might have had different origins. In the Arctic Ocean, only Thuiaria seems to have radiated at higher rates than in the lower latitudes, but the systematic position of most of its representatives is still in need of proper re-evaluation, since they might belong to different genera. Antarctic species previously assigned to Thuiaria are currently in the genus Staurotheca (Peña Cantero et al., 1997). In this sense, Stepanjants et al. (2006) hypothesized a common origin for both genera based on morphology and distributional trends (see below).

Concepts of bipolarity

The idea of bipolarity and its applicability in various polar taxa have been a matter of discussion since the earliest observations by Darwin (1872). After several attempts, it was redescribed as the disjunct distribution of a species (or higher taxa) in cold, temperate and sub-tropical zones of both hemispheres but absent in the tropics (Stepanjants et al., 1997). Later, a second type of bipolarity was defined given the presence of some bipolar taxa in bathyal depths in the tropics, including taxa with an equatorial submergence and thus with a non-disjunct distribution (i.e. maintaining gene flow). A final group of species absent in both tropical shallow waters and both polar regions have been regarded by some authors as bipolar (Stepanjants et al., 2006), though they should be regarded as amphitropical (cf. Allcock & Griffiths, 2015) given their absence in at least one of the poles.

A list of 23 bipolar medusozoan species (of which 15 belong to Leptothecata and Anthoathecata) was provided by Stepanjants et al. (2006). However, most of them were absent from the Antarctic region and should not be considered bipolar (e.g. species belonging to Halopsis, Kirchenpaueria, Nemertesia and Sertularella, among others). This study has identified 16 species inhabiting both polar regions, of which eight are widely distributed and well known in the shallow waters of tropical regions. Among the remaining eight, Abietinaria abietina (Linnaeus, 1758), Grammaria abietina, Obelia longissima (Pallas, 1766), Paragotoea bathybia Kramp, 1942, and Staurostoma mertensii (Brandt, 1834) were already mentioned by Stepanjants et al. (2006). Ptychogena hyperborea Kramp, 1942 and Rhabdoon reesi (Shirley & Leung, 1970) were not mentioned, but the genera they belong to were considered bipolar by these authors. Finally, Plotocnide borealis Wagner, 1885 was not considered by Stepanjants et al. (2006). Some aspects prevent us from considering some of these eight species as bipolar. Ptychogena hyperborea, R. reesi and S. mertensii are Arctic hydromedusae, with only one record from Antarctic waters (see Appendices 2a, 2b). Similarly, A. abietina and P. borealis were only found on one occasion in the shallow waters of South Georgia, i.e. the limit of the Antarctic region. The last three species (i.e. Grammaria abietina, Obelia longissima and Paragotoea bathybia) might be considered putatively bipolar, although all of them are widely distributed and are only lacking in the tropics. Interestingly, Obelia longissima from the high latitudes of both hemispheres are remarkably similar and clearly conspecific (Govindarajan et al., 2006). With the available evidence, there are no true exclusively bipolar Hydroidolina species known.

At a higher taxonomic level, there are some examples of bipolarity in hydrozoans, with polar representatives restricted to cool waters and others that are still pending confirmation. Stepanjants et al. (2006) considered the family Candelabridae to be an example of a bipolar taxon. Species belonging to Monocoryne are cold-water inhabitants in the Arctic, boreal Atlantic and Pacific, and in the southern hemisphere near South Africa (Stepanjants et al., 2006) and Antarctica (Peña Cantero, 2018). However, some species belonging to Candelabrum have also been reported from temperate and sub-tropical shallow waters and deep tropical waters (Segonzac & Vervoort, 1995). This evidence indicates that Candelabrum (and therefore Fam. Candelabridae) cannot be considered a bipolar taxon. Other bipolar genera include Bouillonia, with the two known species restricted to polar regions, one in each (Svoboda et al., 2006), and Gymnogonos, with four total species, two from the North Atlantic and Arctic Seas, one from the North Pacific, and one from the Antarctic region (Stepanjants & Svoboda, 2008). Ptychogena and Rhabdoon are hydromedusae with a few species each, which are mainly distributed in the polar regions and are not reported in the tropics, and therefore probably bipolar (Stepanjants et al., 2006, present contribution). However, the unknown polyp stage and the taxonomic uncertainties of both taxa prevent uncovering their true distribution (e.g. Schuchert, 2010; Schuchert et al., 2017). Similarly, the few species belonging to Margelopsis were mainly found at high latitudes, including the polar regions (see Appendices 1a, 2a), but some records from the tropics still require revision (cf. Kramp, 1961; Stepanjants et al., 2006). Finally, Svoboda et al. (1997) discussed a case of ecological bipolarity in Hydrozoa for two closely related Hydractinia species associated with ophiuroids, which are to date only known from polar waters: Hydractinia ingolfi Kramp, 1932 from the Arctic and Hydractinia vallini Jäderholm, 1926 from the Antarctic region.

The disjunct distribution of the current taxa has been attributed to either prior cosmopolitan species becoming isolated vicariantly in high latitudes during interglacial periods or alternatively, as the result of transequatorial dispersal occurring during glacial maximum cooling (Allcock & Griffiths, 2015 and literature therein). These bipolar divergences have occurred in numerous taxa throughout geological history (Crame, 1993), mainly after the last glaciation, although some examples with an older origin have been suggested (Stepanjants et al., 2006; Allcock & Griffiths, 2015). Population size, maintenance of gene flow and long distance dispersal capabilities across deep waters are crucial factors allowing (or preventing) bipolar speciation (cf. Pawlowski et al., 2007). Up to 230 metazoan marine species could be bipolar (Gutt et al., 2010), including macrobenthic fauna [e.g. the amphipod Eurythenes gryllus (Lichtenstein in Mandt, 1822) (Havermans et al., 2013)] and even sessile benthic brooders with limited dispersal capability [e.g. the bryozoan Callopora weslawski Kuklinski & Taylor, 2006 (Kukliński & Barnes, 2010)]. However, there is evidence of species in different phyla that were previously considered to be bipolar taxa turning out to be different species due to changes in their taxonomic status revealed by molecular techniques (Pabis et al., 2015; Allcock & Griffiths, 2015). Thus, bipolarity at the species level is the exception rather than the rule (Brandt et al., 2012), but it is quite usual at the genus level (e.g. Moles et al., 2017). This might be the case for some bipolar hydrozoan genera, e.g. Bouillonia, Gymnogonos, Monocoryne and Ptychogena. The first three were considered relict taxa by Stepanjants et al. (2006). Further investigations are needed to confirm some of these aspects.

Life history strategy shapes distribution patterns

In agreement with the results for global Hydrozoa obtained by Gibbons et al. (2010), polar representatives with a medusa stage are less represented, display wider distributions and have, consequently, lower endemism levels (Fig. 3). Similar patterns have been reported in polar amphipods (Jażdżewski et al., 1995) and other taxa with meroplanktonic or holoplanktonic representatives (Gibbons et al., 2005). The underlying factors behind this pattern include the general homogeneity of the pelagic environment, the absence of barriers for dispersal, vertical mixing and water flow, and the structure of the food webs (Gibbons et al., 2010). Notably, Arctic meroplanktonic hydromedusae are more widely distributed than they are in the Antarctic, and subsequently, endemism in the Southern Ocean is higher. Walkusz et al. (2004) found a similar trend when comparing Arctic and Antarctic neritic mesozooplankton. Our findings thus confirm the effectiveness of the Antarctic Polar Front as a barrier (or filter) for dispersal, as has been widely suggested (e.g. Crame, 1999; Clarke et al., 2005). However, there are other polar groups, such as polychaetes, in which long-lasting planktonic stages do not correlate with larger geographical ranges (Pabis et al., 2015). Remarkably, opposite results were found in boreal (Cornelius, 1981) and Mediterranean Sea (Boero & Bouillon, 1993) hydrozoans. The latter study reported no contrasting differences between the number of medusa-producing and medusa-lacking species in the entire dataset or when considering the endemic species separately. These authors noticed that the widely distributed taxa mainly included species without a medusa stage. The reason behind these different findings might be the inability to discriminate the taxa using current taxonomy based on morphological species identification, resulting in the globally cited taxa, as has already been noted by Boero & Bouillon (1993). A growing number of studies using diverse genetic techniques have recently shown that several widespread hydroid species are actually species complexes (e.g. Schuchert, 2014; Postaire et al., 2016, 2017). The wide distribution of some Hydrozoa species is thus a taxonomic artefact.

Missing species from the polar regions and recent discoveries

To provide an updated biodiversity inventory, especially in the light of global warming effects, both the latest data on new species discoveries and knowledge on missing or extinct species must be incorporated. The latter is particularly difficult to assess, as it is open to question how much time is needed since the last species record to attribute a status of missing or extinct. According to Gravili et al. (2013), who found 67 (17%) hydrozoan species with no positive records in the last 40 years from the Mediterranean Sea, “four decades of absence might be long enough to suggest the possibility that regional extinction (or at least range contraction) may be considered”. However, the differences in sampling and taxonomic efforts between the well-studied Mediterranean Sea and the less-known polar regions force us to follow the proposed 40-year period with caution.

Among the species considered rare, there is a subset of species whose presence in the polar regions has not been confirmed in the last 40 years [89 (38%) in the Arctic and 64 (25%) in the Antarctic]. In addition, several very rare taxa have been “missing” for the last 100 years. Among them, some are only known from their original description and undoubtedly do not belong to any other species described [e.g. the Antarctic Halecium pallens (Peña Cantero, 2014b) and Rhysia halecii (Peña Cantero, 2015)]. Other rare species still require taxonomic validation, as they might be conspecific with better known species (e.g. the hydroids Eudendrium cyathiferum, Tubularia antarctica and Sertularella nuttingi, and the hydromedusa Bythotiara drygalskii, whose polyp is unknown). A few species are widely distributed but have been recorded rarely in polar regions, and their presence has not yet been confirmed [e.g. Abietinaria abietina was reported only once from South Georgia]. Finally, some others are well known, widespread taxa reported in polar regions, most likely as a result of misidentification due to incomplete taxonomic knowledge. This is the case for Eudendrium ramosum (Linnaeus, 1758), which is probably absent from Antarctic waters, whose records still need reconfirmation. In this regard, the globally distributed Halecium delicatulum Coughtrey, 1876 was considered to occur in Antarctic waters until the recent work by Peña Cantero (2014b), who demonstrated that the previous records corresponded to a different species.

The discrepancy in the proportion of rare and missing species between life cycle strategies could be biased by differences in the sampling and taxonomic effort rather than real ecological dissimilarities. In this regard, it is worth mentioning that many Antarctic benthic hydrozoans have been sampled and reported in the last 40 years. In Arctic waters, however, the proportion is considerably lower (Fig. 5). The opposite patterns are observed for the meroplanktonic taxa (Fig. 5). Analogously, Antarctic hydromedusae and Arctic benthic hydroids are the groups with a higher proportion of rare taxa (Fig. 4). The benthic hydrozoans from the Antarctic waters are more comprehensively studied than the planktonic taxa. Many new Antarctic benthic hydroids have been described during the course of this century (Xavier et al., 2013), but this is not the case for Arctic waters. In the Arctic, despite being less obvious, the situation might be the reverse, with a better knowledge of the planktonic taxa than the benthic taxa. Therefore, given the differences in sampling effort mentioned above and the absence of long-term studies of polar Hydrozoa, we consider that there is not enough evidence to claim that regionally extinct species exist.

Contrasting bathymetric patterns

According to the present results, the hydrozoan bathymetric distribution mirrors the geomorphological characteristics of each region. The bathymetric differences in shelf break occurrence between both polar regions are responsible for the uneven distribution with depth of polar Hydrozoa.

The continental shelf in the Arctic is mostly shallow, with the largest area between 0 and 50 m depth (excluding deeper regions of the Barents Sea, Beaufort Sea and northern Greenland), making up 22% of the entire Arctic Ocean (Jakobsson et al., 2012), while in the Antarctic, the continental shelf is very narrow and deep (Dayton, 1990). For this reason, hydrozoan species richness is dissimilar when comparing depth ranges (Fig. 6) but similar when comparing zones: the number of species present in both continental shelves is not markedly different (202 in the Arctic vs. 219 species in the Antarctic). The hydrozoan species occurrence with depth from several locations displays a unimodal pattern, with maximum values at approximately 100 m depth (Altuna, 2007 and literature cited). In the poles, the species occurrence with depth is also unimodal, and the maximum number of species occurs deeper than in other regions located at lower latitudes, probably due to the deeper shelves and disturbance by ice scouring (cf. Mercado Casares & Peña Cantero, 2018; present study). Thus, as expected, the results differ between polar regions: the Arctic harbours a greater number of species between 30 and 200 m, while in the Antarctic, the greatest number of species occurs at depths of 200 and 500 m, with both regions having the shelf as the most speciose bathymetric range (Fig. 6). Specifically, the intermediate sectors of the shelf in the Antarctic harbour a higher number of species, which agrees with the results of Mercado Casares & Peña Cantero (2018).

The shallowest areas are unevenly represented, with the number of species in the Arctic more than twice the number found in the Antarctic (Fig. 6). The group of shallow water hydrozoans in the Antarctic shows low similarity with any other bathymetric group considered, with approximately 20% of the species exclusively known from shallow waters. Conversely, the Arctic shallow water group is clustered with the following depth range (Fig. 8). The noticeable lower hydrozoan diversity and higher dissimilarity in Antarctic shallow ecosystems might be simply a result of a lower sampling effort in this zone (Peña Cantero et al., 2013). Shallow areas are extremely difficult to sample in high Antarctic regions due to the permanent cover of coastal and nearshore regions by fast ice and grounded glacier ice (Dimmler et al., 2001). The strong effects of ice scouring and anchor ice in Antarctic waters might be other factors to consider resulting in the differences reported (Gutt et al., 1996; Conlan et al., 1998). In addition, kelp forests in the Arctic act as ecosystem engineers and habitat formers, creating hotspots of associated invertebrate diversity in shallow regions (Włodarska-Kowalczuk et al., 2009), with many hydrozoan species using their rhizoids and fronds as substrates (Ronowicz et al., 2011). Kelp forests are, however, absent from the Southern Ocean (Santelices, 2007).

The present findings show wider depth ranges in the Antarctic, regardless of the range considered, but the differences are not very pronounced. When considering those species with depth ranges above 300 m [i.e. 138 (56%) from Antarctic and 112 (50%) from Arctic waters], most of the representatives from each region may be classified as eurybathic (following Menzies et al., 1973). These patterns are congruent with the extended eurybathy reported for other Antarctic marine invertebrates, such as polychaetes, mollusks, crustaceans (Brey et al., 1996) and echinoderms (Martín-Ledo & López-González, 2014) as well as for the Arctic deep-water meiofauna, macrofauna and megafauna whose depth ranges stretch into the shelves (Bluhm et al., 2011). The extended eurybathy in the benthic Arctic and Antarctic invertebrates has been attributed to contraction–expansion faunal movements between the slope and refugia that took place between the glacial and interglacial periods during the Pleistocene (Brey et al., 1996; Bluhm et al., 2011; Gili et al., 2016). On the other hand, Jumars & Fauchald (1977) suggested that depth ranges of Antarctic species should exceed those of Arctic species given that food availability is greater in the former. Bilyard (1991) tested this hypothesis using polychaetes and obtained weak support. The present results also show little evidence.

Previous studies dealing with Antarctic hydrozoans have emphasized the great eurybathy within the group (Stepanjants, 1979; Peña Cantero & Garcia-Carrascosa, 1999; Peña Cantero 2004, 2014a; Mercado Casares & Peña Cantero, 2018). Interestingly, given that many supraspecific taxa occur at great depths, the colonization of deep waters would have occurred early in the evolution and diversification of Hydrozoa (Fernandez & Marques, 2018). In contrast, Fernandez & Marques (2018) reported reduced eurybathy for Antarctic species compared to other Atlantic regions, including the Arctic. These authors specified that no Antarctic species has bathymetric ranges > 1500 m, with only three species reported below 1000 m depth. However, our results show 21 species with ranges above 1500 m and 50 species recorded below 1000 m. Fernandez & Marques (2018) centred their analyses in the corresponding Atlantic sector and on records deeper than 50 m. Thus, the exclusion of the shallow water records and most of the Antarctic region and the inclusion of the widely distributed meroplanktonic hydromedusae in our study explain the narrower ranges obtained by these authors. In any case, according to Fernandez & Marques (2018), polar hydrozoans display smaller and less variable ranges than their sub-tropical and tropical counterparts.

Conclusions

Using the hydrozoan fauna as a model group, the results confirm some previous hypotheses about marine life in polar waters. Thus, the species richness of Hydrozoa (Leptothecata and Anthoathecata) is only slightly lower in the Arctic compared with the Antarctic. A low number of species per genus seems to be a characteristic feature of polar hydrozoans, with only a few prolific genera in both regions. The Arctic and Antarctic regions differ in their rates of endemism. The high endemism found in Antarctic Hydrozoa is a reflection of the complete isolation of the Antarctic since the Oligocene, while the low endemism in the Arctic reflects the linkage of the Arctic Ocean with the Atlantic and Pacific Oceans. The class Hydrozoa is well represented in the polar regions. The PR value (the contribution of polar representatives to the total known species within a group) of Hydrozoa is among the highest when compared to other invertebrate taxa. The life history strategy of polar hydrozoans is characterized by the predominance of benthic taxa reproducing by fixed gonophores over taxa releasing a free-swimming medusa. The majority of species with an effective, long-living dispersal stage (i.e. a medusa) have wide distributions in both regions. In contrast, most of the benthic species in the Antarctic are endemics, but benthic species in the Arctic usually have a wide distribution. Bipolarity in hydrozoans is attributed to three species (i.e. Grammaria abietina, Obelia longissima and Paragotoea bathybia), although all of them occur worldwide, with the exception of the tropics. Therefore, bipolar Hydrozoa species de facto do not exist.

The taxonomic effort (efforts to identify individuals to the species level) in this particular group is more likely higher in the Antarctic, especially in recent times. Species new to science have been discovered more often in the Antarctic in recent years (since 2000). There are several extremely rare species in both polar regions that have not been recorded for the last 100 years (16 species from each polar region). Some of them might be considered putatively extinct, but they are more likely taxonomic artefacts.

The Arctic harbours the greatest species diversity between 30 and 200 m, while in the Antarctic, the greatest species diversity is found between 200 and 500 m, with both regions having the shelf as the most speciose bathymetric range. The bathymetric trends of polar Hydrozoa are shaped by their continental shelf geomorphologies.