Polar Biol (2014) 37:859–877 DOI 10.1007/s00300-014-1487-9 ORIGINAL PAPER Diversity, abundance and composition in macrofaunal molluscs from the Ross Sea (Antarctica): results of fine-mesh sampling along a latitudinal gradient Stefano Schiaparelli • Claudio Ghiglione • Maria Chiara Alvaro • Huw J. Griffiths • Katrin Linse Received: 2 August 2013 / Revised: 11 March 2014 / Accepted: 12 March 2014 / Published online: 26 March 2014  Springer-Verlag Berlin Heidelberg 2014 Abstract The Latitudinal Gradient Program (2002–2011) aimed at understanding the marine and terrestrial ecosystems existing along the Victoria Land coast (Ross Sea), an area characterized by strong latitudinal clines in environmental factors. During the program’s voyage of the Italian RV ‘‘Italica’’ in 2004, a fine-mesh towed gear, the ‘‘Rauschert dredge’’, was deployed for the first time at 18 stations in four latitudinal distinct shelf areas between *71S and *74S. The collected samples contained undescribed species and new records for the Ross Sea from a variety of different marine taxa. Here, we describe the molluscan fauna and investigate evidences for latitudinal effects on molluscan diversity, abundance and assemblage composition. No significant latitudinal trends were detected: while diversity did not vary significantly with latitude, species richness showed an apparent but non-significant decrease with increasing latitude. Beta-diversity was found to be high both within and between latitudinally distinct shelf areas. A large fraction (*20 %) of the collected molluscs corresponded to new species records for the Ross Sea or undescribed species. Rarity in Antarctic molluscan occurrences was confirmed, with singletons (i.e. species represented by only a single individual) accounting for a 22 % and uniques (i.e. species occurring in one sample S. Schiaparelli (&)  C. Ghiglione  M. C. Alvaro Department of Earth, Environmental and Life Sciences (DISTAV), University of Genoa, Genoa, Italy e-mail: stefano.schiaparelli@unige.it S. Schiaparelli  C. Ghiglione  M. C. Alvaro Italian National Antarctic Museum (MNA) (Section of Genoa), University of Genoa, Genoa, Italy H. J. Griffiths  K. Linse British Antarctic Survey, Cambridge, UK only) for a 43.5 % of the total presence. Our study of the smaller macrofaunal benthic fraction showed that Antarctic marine research still has far to go to have robust reference baselines to measure possible changes in benthic communities, even in the case of the assumed well-known, wellsampled and well-studied group of Ross Sea shelf molluscs. We advocate the use of fine-mesh trawling gears for routine sampling activities in future Antarctic expeditions to assess the full marine biodiversity. Keywords Antarctica  Benthos  Mollusca  Ross Sea  Victoria Land coast  Rauschert dredge Introduction In Antarctica, the Ross Sea sector is one of the better-known areas due to the high number of scientific expeditions that have been undertaken there since 1899, when during the Southern Cross Expedition (1898–1900) men overwintered for the first time on the Antarctic continent. This challenging adventure opened the Ross Sea sector to a long series of other scientific expeditions (i.e. the National Antarctic Expedition, 1901–1904; the British Antarctic Expedition, 1907–1909; the British Antarctic Expedition, 1910–1913; the R.S.S Discovery II, 1935–1937 and 1937–1939; the U.S Navy Antarctic Expedition, 1947–1948; the Deep Freeze I, 1955–1956; the Deep Freeze II, 1956–1957; the Deep Freeze III, 1957–1958; the Deep Freeze IV, 1958–1959; the Trans Antarctic Expedition, 1956–1958; the Stanford University Invertebrate Studies, 1958–1961; the New Zealand Oceanographic Institute, 1958–1960; the USAP Project, 1960–1961; the USAP Eltanin 27, 1967; and the USAP Eltanin 32, 1968), which collected an impressive number of scientific samples. 123 860 After this initial phase of geographical and biological exploration, including species description, there was a progressive shift in research focus from large-scale biological sampling, aimed at species inventories and descriptions, to the study of community assembly and ecosystem functioning at the end of the 1960s. This change was partly due to an increase in interest for ecological disciplines, but also to the establishment of permanent bases: McMurdo Station (built in 1955–1956) and Scott Base (built in 1959), both located at *78S, and Mario Zucchelli Station (built in 1985) located at *74S which enabled sampling on a regular basis and ecological studies related to the seasonal changes in environmental conditions and the dynamics occurring at high latitudes (e.g. Dayton et al. 1970; Dayton and Oliver 1977). Thanks to all these recent and historical scientific activities, resulting in a great number of stations sampled in this region, the shelf fauna of the Ross Sea is considered to be one the better sampled in the Southern Ocean (Clarke et al. 2007; Griffiths et al. 2011). Within this region and in general for Antarctica, one of the best-studied taxa is the Phylum Mollusca, which is only second to Arthropoda in terms of number of collection records (De Broyer and Danis 2011). In fact, virtually all scientific expeditions to the Ross Sea collected molluscan specimens, forming an impressive amount of scientific material which has accumulated on museum’s shelves decade after decade since 1899. In 1990, Dell, after having sorted, organized and classified a large part these mollusc samples, in an unprecedented effort which lasted for about 30 years (Dell 1990:1–2), published the volume ‘‘Antarctic Mollusca’’ (Dell 1990), a synthesis of the taxonomical and distributional knowledge of Antarctic molluscs, which embraced almost all available samples collected at least until the end of the 1970s. This book had a special focus on the Ross Sea, due to the great number of samples collected in this area and listed 193 species of shelled Mollusca, out of which 43 were described as new (Dell 1990). In later years, only few adjustments were made to the number of species known for the Ross Sea, which increased to 197 species (Griffiths et al. 2009), roughly corresponding to *30 % of the total number of mollusc species (n = 684) known for the Southern Ocean (Griffiths 2010). All the above-cited sampling activities were performed using a variety of sampling gears, ranging from grabs to trawl nets, which should have assured the sampling of all size classes. However, as Dell (1990) clearly stated in his contribution, in all the studied Antarctic samples, there always was an apparent lack of ‘‘microfauna’’ or smallsized macrofauna. Dell explained this was artefact of the design of sampling activities and gears used, where most of 123 Polar Biol (2014) 37:859–877 the sediment was probably washed away together with the small-sized species (Dell 1990:264). This was the state of the art until 2004, when a joint Italian-New Zealand ship-based, large-scale sampling campaign was performed in the Ross Sea under the Latitudinal Gradient Program framework (LGP, http://www. lgp.aq/). Previously, for almost 40 years, no large-scale benthic sampling activities occurred in the area and most sampling was near shore oriented in the vicinity of permanent research stations. The LGP was an international research project running from 2002 to 2011, which aimed at understanding of the complex ecosystems that exist along the Victoria Land coast and the effects of environmental change on these ecosystems (http://www.lgp.aq/). Along the natural gradient of the Victoria Land coast, in fact, environmental factors such as solar radiation, temperature and sea-ice cover vary with latitude (Berkman et al. 2005) and are likely to affect the diversity and distribution of benthic communities. Other environmental gradients such as depth below sea level, height above sea level and distance from the coast complicate the latitudinal gradient influence on the coastal ecosystems (Cummings et al. 2010 and references therein). All these factors and gradients make the Victoria Land coast one of the few natural latitudinal gradients where it is possible to investigate the potential effect of environmental parameter gradients on the benthic communities. In 2004, two parallel LGP voyages, i.e. RV ‘‘Tangaroa’’ TAN0402 BioRoss and ‘‘Italica’’ ‘‘Latitudinal Gradient Program’’, sampled along the Victoria Land coast collecting new records of molluscs from the Ross Sea which accounted for a *20 % of the total of species found (Schiaparelli et al. 2006). During the RV ‘‘Italica’’ expedition, the fine-meshed ‘‘Rauschert dredge’’ was deployed for the first time on the shelf of the Ross Sea (Rehm et al. 2006). The ‘‘Rauschert dredge’’ (Lörz et al. 1999), like the ‘‘Brenke-sled’’ (Brenke 2005), belongs to the category of ‘‘fine-mesh towed gears’’, i.e. those having mesh sizes B500 lm and are specifically designed to retain epibenthic specimens of small size. The use of fine-mesh towed gears has progressively increased in recent years, especially in the Atlantic sector of the Southern Ocean (e.g. Linse 2006; Schrödl et al. 2011), allowing a considerable increase in the knowledge of species distribution and occurrence, especially for deep-sea taxa (Kaiser et al. 2013 and references therein), but had never previously been used in the Ross Sea. More recently, a ‘‘Brenke-sled’’ was used in the Ross Sea during the RV ‘‘Tangaroa’’ TAN0802 in 2008 (Lörz et al. 2013). To date, several papers have been published based on the material collected by this dredge during the 2004 RV ‘‘Italica’’ research voyage. In general, they focus on Polar Biol (2014) 37:859–877 861 Peracarida (Rehm et al. 2007), with specific studies on Isopoda (Choudhury and Brandt 2007, 2009; Choudhury et al. 2011) and Cumacea (Rehm and Heard 2008), including descriptions of few species and subspecies. Molluscs, although not being numerically dominant in the ‘‘Rauschert dredge’’ samples (they accounted only for a 9.6 % in abundance and 4.2 % in biomass; Rehm et al. 2006), were represented mostly by small-sized specimens, i.e. to the size fraction that was considered by Dell (1990) to be ‘‘lacking’’. This small macrobenthic fraction had a great potential of including new species records as well as undescribed species. Following the targets of the Latitudinal Gradient Program, here, we present the small macromolluscan fauna of the Victoria Land Area and investigate any effect of latitude on this group for diversity, abundance and assemblage composition. Materials and methods Study area and sample processing In the Austral summer 2004, during the 19th PNRA Antarctic expedition on board, the RV ‘‘Italica’’, a ‘‘Rauschert dredge’’, was deployed between Cape Adare (*71S) and Terra Nova Bay (*75S) at eighteen stations with depths varying from 84 to 515 m (Fig. 1; Table 1). The ‘‘Rauschert dredge’’ was equipped with a net of 500-lm mesh size and an opening width of 0.5 m (Lörz et al. 1999). The dredge was towed at a mean velocity of 1 knot (1,852 km/h) and calculated haul distances ranged from 59 to 575 m (Table 1 and Rehm et al. 2006). The collected material was sieved on board through a 500-lm mesh, preserved in 90 % pre-cooled ethanol and kept at -25 C for later DNA extraction. A DNA barcoding activity on this material is currently underway within the research program BAMBi (Barcoding of Antarctic Marine Biodiversity, 2010/A1.10) funded by the Italian National Antarctic Research Program (PNRA) (Ghiglione et al. 2013). In the laboratory, all living specimens were sorted under a stereomicroscope and, whenever possible, classified down to the species level. Minute species were photographed using an ESEM (Leo Stereoscan 440). Dead shells, i.e. the thanatocoenosis, were not taken into account in the present study, which refers only to live collected specimens. Data analysis Species richness was evaluated within and between each of the four latitudinal sampling areas (i.e. *71S, *72S, *73S and *75S), called latitudinal bins from here on. Firstly, individual-based interpolation (rarefaction) and Fig. 1 Victoria Land coast where the 2004 RV ‘‘Italica’’ voyage took place with the sampled areas of Cape Adare, Cape Hallett, Coulman Island and Cape Russell. Position of sampling stations is marked with black circles. Stations’ coordinates are reported in Table 1 extrapolation curves for all the stations sampled at each of the four latitudinal bins considered were compared, using the method proposed by Colwell et al. (2012). Secondly, species richness was compared between latitudinal bins by combining together all the samples within a bin and calculating sample-based interpolation (rarefaction) and extrapolation curves, based on incidence data from reference samples (Colwell et al. 2012). Uncertainty of estimations is reported in terms of unconditional confidence intervals (at 95 %) under the multinomial model (for Individual-based interpolation/extrapolation curves) or under the Bernoulli product model (for the sample-based interpolation/extrapolation curves) (Colwell et al. 2012). The non-overlap of 95 % confidence intervals was used as the indication of statistical difference (Colwell et al. 2012; Gotelli and Ellison 2013). Rarefaction and extrapolation analyses were conducted using the on-line resource iNEXT (Hsieh et al. 2013), which is based on R-statistical language (http://www.r-project.org). In sample-based rarefaction analysis, the Coulman Island data were not included 123 862 Polar Biol (2014) 37:859–877 Table 1 Rauschert dredge stations of the Victoria Land Transect Cruise (Ross Sea, Antarctica) Station Date North–South Position Latitude (S) Longitude (E) Depth (m) Haul length (m) Sediment Cape Adare A1 15/02/2004 7115.50 17041.90 515 358 Sand with few pebbles and stones A2 14/02/2004 7117.30 17039.20 421 298 Sand and gravel A3 14/02/2004 7118.7 0 0 17029.2 305 257 Sand A4 14/02/2004 7118.40 17028.90 230 376 Sand and pebbles A5 15/02/2004 7118.70 17025.50 119 59 Cape Hallett Hout1 09/02/2004 7215.70 17024.80 458 375 Mud and pebbles Hout2 11/02/2004 7217.5 0 0 17029.4 353 375 Sandy mud and stones Hout4 12/02/2004 7218.50 17026.80 235 194 Sand Hin2 10/02/2004 7216.90 17012.20 391 186 Coarse sand and small gravel Hin3 16/02/2004 7217.00 17013.10 316 194 Muddy sand with stones Hin4 16/02/2004 7217.10 17014.00 196 169 Mud and sand 16/02/2004 7217.2 0 0 17017.9 84 113 Small gravel 18/02/2004 7324.50 17023.20 474 375 Mud and small gravel 18/02/2004 7322.7 0 17006.90 410 153 Mud and pebbles 20/02/2004 7443.20 16413.10 366 192 Sand with gravel and stones R2 21/02/2004 7449.0 0 0 16418.1 364 575 Fine sand R3 20/02/2004 7449.30 16411.50 330 565 Rock, sand, mud and pebbles R4 20/02/2004 7449.30 16411.50 208 97 Hin5 Sand with pebbles and stones Coulman Island C1 C2 Cape Russell SMN Rock, mud and large stones Sediment data from Rehm et al. (2006) in calculations due to the inadequacy of the sample size (i.e. only two sites were available), following Gotelli and Ellison (2013, 467). Beta-diversity (i.e. species turnover) between samples and between latitudinal bins was evaluated by considering incidence data and partitioning pairwise gamma diversity into additive components by following the SDR simplex methodology (Podani and Schmera 2011). The additive absolute components of gamma diversity for sites (species shared, beta-diversity or species turnover, richness difference, richness agreement, species replacement, nestedness) were graphically reported in ternary plots (e.g. Fig. 5), and their percentage values reported in tables (e.g. Table 4). The above analyses were performed with the software SYN-TAX 2000, and ternary plots were produced using its module ‘‘Non-hierarchical clustering’’ (Podani 2001). The program PRIMER 6 (Plymouth Marine Laboratory) (Clarke and Gorley 2005) was used to calculate diversity indices. For each sampling station, we calculated: the abundance (N), the number of species (S), the ShannonWiener diversity index (H0 ) (in log10) and the Pielou’s evenness (J0 ). Although Shannon-Wiener H0 is not the best performing available index for measuring biodiversity (Buckland et al. 2011), we preferred to keep using this one 123 in order to allow comparisons with published literature (e.g. Rehm et al. 2007, cfr. Table 5 and Schiaparelli et al. 2006, cfr. Fig. 2). H’ and Simpson’s values were averaged per latitudinal bins and compared by one-way parametric ANOVA. Homogeneity of variances was tested with Bartlett’s test and normality checked with the Shapiro test. When necessary, a post hoc Tukey’s HSD test was used to evidence differences between latitudes. Post hoc comparisons were made using the ‘‘agricolae’’ package in R. Abundance and composition Due to the different lengths of hauls, number of individuals were standardized to 1,000 m2 hauls (Table 2; Fig. 7) according to (Rehm et al. 2007) in order to allow comparisons between hauls and with published literature (e.g. numbers of Peracarida: Table 2 in Rehm et al. 2007). Possible differences between groups of stations, due to depth and latitude, were studied with multivariate techniques. Abundances in each sample were standardized by total and then transformed by square root to down weight the contribution of most abundant species, or, alternatively, presence/absence data were used. Bray-Curtis similarity coefficients were calculated between stations and displayed Polar Biol (2014) 37:859–877 863 Fig. 2 Box-plot of H’ and Simpson’s diversity indices at the four considered areas by non-metric multidimensional scaling (nmMDS). Oneway ANOSIM was then used to test for the factor ‘‘depth’’, with pre-defined levels: ‘‘1’’ (0–100 m), ‘‘2’’ (101–200 m), ‘‘3’’ (201–300 m), ‘‘4’’ (301–400 m), ‘‘5’’ (401–500 m), ‘‘6’’ ([501 m) and the factor ‘‘latitude’’, with subjectively chosen levels: ‘‘1’’ (stations between 71S and 72S), ‘‘2’’ (between 72S and 73S), ‘‘3’’ (between 73S and 74S), ‘‘4’’ (between 74S and 75S). In the multivariate analyses, we excluded 14 specimens, which were not determinable due to the small dimensions or the lack of morphological characters (Tritonia spp., Toledonia sp. juv., Pseudokellya sp. juv., Opisthobranchia indet.). All Solenogastres were sorted into 7 OTUs (operational taxonomic units) and included in this form in the analyses. Species, which were not classified to the specific level, were indeed included in the multivariate analyses and reported at the level of genus or family. Multivariate analyses were performed with the software PRIMER 6 (Plymouth Marine Laboratory) (Clarke and Gorley 2005). Taxonomy The choice of the right specific name, for several taxa, has been problematic due to the fact that, for several groups, there are no recent taxonomic revisions available. Moreover, there is an apparent state of flux at the family level for several taxa, following results from recent molecular studies (e.g. Bouchet et al. 2011) and different interpretations given by different authors (e.g. Engl 2012). In the latest comprehensive review of Antarctic Mollusca (Engl 2012), for example, there are several affiliations at the family level, which do not reflect those currently in use and reported in the World Register of Marine Species (WORMS) (www.marinespecies.org, last search 30 July 2013). In Engl (2012), for example, the genera Brookula Iredale, 1912; Lissotesta Iredale, 1915; and Leptocollonia Powell, 1951 are all placed in Turbinidae, a family characterized by heavily calcified opercula. Of these species, only Leptocollonia has a calcareous operculum and belong to Collonidae (as correctly reported in WORMS), while Brookula and Lissotesta are considered Seguenzioidea with unassigned family in WORMS and do not have a calcareous operculum. This instability at the family would have made, in many cases, any choice of a family assignment arbitrary. Due to this uncertainty, in this study, we have decided to refrain from any taxonomic revision. For supraspecific classifications, we adhered to WORMS while, for the specific affiliations, we have followed (unless differently stated in notes) the classification of Engl (2012), since it is based in most cases on direct comparison with type materials. The classes Gastropoda, Bivalvia and Monoplacophora were classified by S. Schiaparelli; Solenogastres, Polyplacophora and Scaphopoda by K. Linse. The complete list of 123 864 Polar Biol (2014) 37:859–877 Table 2 Abundance of Mollusca along the Victoria Land coast (Ross Sea, Antarctica) Station Gastropoda N Bivalvia N*1,000 m -2 N Monoplacophora N*1,000 m -2 N -2 N*1,000 m Solenogastres Polyplacophora -2 N N*1,000 m N N*1,000 m Scaphopoda -2 N N*1,000 m-2 Cape Adare A1 18 100.56 0 0 0 0 5 27.93 0 0 0 0 A2 70 469.80 64 429.53 0 0 39 261.74 0 0 0 0 0 A3 515 4,007.78 26 202.33 0 0 15 116.73 0 0 0 A4 3,885 20,664.89 132 702.13 0 0 46 244.68 0 0 0 0 A5 58 1,966.10 6 203.39 0 0 1 33.90 0 0 0 0 0 41 218.67 0 0 18 96 0 0 2 0 0 Cape Hallett Hout1 0 10.67 Hout2 18 96 35 186.67 3 16.00 39 208 0 Hout4 197 2,030.93 63 649.48 35 360.82 383 3,948.45 4 41.24 0 0 Hin2 76 817.20 0 0 0 0 21 225.81 1 10.75 4 43.01 Hin3 403 4,154.64 173 1,783.51 0 0 99 1,020.62 2 20.62 1 10.31 Hin4 398 4,710.06 61 721.89 0 0 69 816.57 1 11.83 0 0 Hin5 57 1,008.85 91 1,610.62 0 0 136 2,407.08 1 17.70 0 0 0 Coulman Island C1 1 5.33 34 181.33 0 0 0 0 0 0 15 C2 81 1,058.82 467 6,104.58 0 0 29 379.08 0 0 51 666.67 80 770.83 11 114.58 0 0 34 354.17 0 0 0 0 17.39 Cape Russell SMN 74 R2 28 97.39 86 299.13 0 0 5 R3 31 109.73 17 60.18 0 0 0 55 5,965 1,134.02 16 1,323 329.90 0 38 0 10 949 R4 Total 1 3.48 1 3.48 0 0 0 0 0 206.19 0 10 0 0 74 0 N = number of specimens collected per station; N*1,000 m-2 = number of specimens per station standardized to 1,000 m2 species and their distributional data occurrences (available through ANTOBIS, the geospatial component of SCARMarBIN) are reported in Ghiglione et al. (2013). Results Diversity Almost, all the species collected by the ‘‘Rauschert dredge’’ were very small (generally not exceeding 3 mm) and belong to families that are known to include minute species. Only very few individuals were juveniles of species which attain larger sizes when adult (e.g. Limopsis marionensis). From the 18 sampled stations, 8,359 living specimens of molluscs belonging to 161 species were obtained (Table 2). These comprised of 5,965 specimens of Gastropoda (113 species; 72 % of the total species richness), 1,323 specimens of Bivalvia (36 species; 22.4 % of the total), 949 specimens of Solenogastres (7 species; 4.3 % of the total), 74 specimens of Scaphopoda (3 species; 1.9 of the total), 38 specimens of Monoplacophora (1 species; 0.6 % of the 123 total) and 10 specimens of Polyplacophora (1 species; 0.6 % of the total). Thirty-one species (3 Bivalvia, 7 Solenogastres, 1 Polyplacophora, 2 Scaphopoda, 18 Gastropoda corresponding to a total of 1,090 specimens) were assigned to the genus or family level only for a variety of reasons and were treated as OTUs (operational taxonomic units). Thirty-five species were present as singletons (21.74 % of the total number of species), i.e. found with a single specimen, and 13 species as doubletons (8.07 % of the total), i.e. found with two specimens only. Seventy species (43.48 % of the total) were uniques, i.e. occurring in a single station only, and 35 (21.74 % of the total) were duplicates, i.e. occurring at two stations only. Diversity indexes (H0 log10) showed a general constancy across the sampling stations, with values comprised between 0.7 and 1.14 (Table 3; Fig. 2). Parametric ANOVA showed no statistically significant differences between group means, for the considered latitudinal bins, of H0 values (F (3, 14) = 1.54, p = 0.247) while, for Simpson’s index, parametric ANOVA indicated the existence of statistically significant differences (F (3,14) = 3.75, p = 0.036). Post hoc comparisons (Tukey’s HSD test Polar Biol (2014) 37:859–877 865 Table 3 Diversity equitability indexes of the 18 Rauschert dredge stations: S = Total species, N = Total individuals, d = Margalef’s index, J0 = Pielou’s evenness, H0 = Shannon’s index, 1 - k0 = Simpson’s index Sample S N d J0 H0 (log10) 1 - k0 A1 8 128 1.442 0.8882 0.8021 0.8116 A2 24 1,161 3.259 0.763 1.053 0.8701 A3 29 4,327 3.344 0.6146 0.8988 0.805 A4 54 21,612 5.31 0.55 0.9528 0.8107 A5 19 2,203 2.338 0.8549 1.093 0.8899 Hout1 12 325 1.902 0.7737 0.835 0.7942 Hout2 Hout4 18 51 507 7,031 2.73 5.645 0.8107 0.6415 1.018 1.095 0.8686 0.8119 Hin2 23 1,097 3.143 0.8278 1.127 0.8966 Hin3 52 6,990 5.761 0.7399 1.27 0.9099 Hin4 30 6,260 3.317 0.6189 0.9142 0.7599 Hin5 25 5,044 2.815 0.5657 0.7908 0.7335 C1 12 267 1.969 0.6839 0.7381 0.7267 C2 50 8,209 5.437 0.6172 1.049 0.751 SMN 20 1,240 2.668 0.7758 1.009 0.8544 R2 28 421 4.468 0.7627 1.104 0.8598 R3 17 170 3.116 0.9059 1.115 0.9125 R4 33 1,670 4.312 0.8962 1.361 0.9364 with significance level at 0.05) indicated that the mean value of Simpson’s diversity index was not statistically different between Cape Adare and Cape Hallett, but differed from those of Coulman Island and Terra Nova Bay. Species Richness evaluated through individual-based interpolation (rarefaction) and extrapolation curves (Fig. 3) showed clusters of curves having largely overlapping unconditional confidence intervals (at 95 %) at Cape Adare, Cape Hallett and Terra Nova Bay areas. Only at Coulman Island and Terra Nova Bay area did some of the sampled stations showed markedly different values (i.e. with not overlapping confidence intervals) of species richness. At Coulman Island, station C2 almost doubled the richness of station C1, while at Terra Nova Bay, station R4 did not overlapped with the cluster formed by the others. Sample-based interpolation (rarefaction) and extrapolation curves for incidence data showed a cluster of curves with overlapping 95 % confidence intervals (Fig. 4), which indicate that the three latitudes do not statistically differ. It is worth remembering, however, that the number of samples available per latitudinal bin is limited and only Cape Adare–Cape Hallett areas satisfy the recommended threshold (i.e. the availability of at least 5 samples) for this kind of analyses (Gotelli and Ellison 2013, 467). Beta-diversity, calculated as the sum of species replacement and richness difference (Podani and Schmera 2011), was found to be high at all sites but that of Terra Nova Bay, the area also having the lowest species richness (although not statistically significant, Fig. 4), where beta-diversity and richness agreement appear to be more balanced (Fig. 5; Table 4). At all latitudes, with the exception of Coulman Island, only two sampling stations were performed (and thus the result are less informative), the similarity always showed percentage values lower than those of species replacement or richness difference (Table 4). Abundance and composition The relative abundance of species within the six mollusc classes varied at each sampling station. In general, Gastropoda and Bivalvia are dominant, while at same places Solenogastres represented *50 % of the number of species present, e.g. stations Hout4 and Hin5 (Fig. 6). Out of the 161 species found, two exceeded 1,000 individuals (in non-standardized data), i.e. Onoba turqueti (Lamy, 1905) (1,668 individuals) and Powellisetia deserta (E.A. Smith, 1907) (1,060), while other 12 species were found with more than 100 individuals which are, in order of decreasing abundance: Eatoniella kerguelenensis (E.A. Smith, 1907) (971 ind.), Solenogastra sp. 3 (701 ind.), Thyasira debilis (Thiele, 1912) (568 ind.), Anatoma euglypta (Pelseneer, 1903) (510 ind.), Onoba gelida (E.A. Smith, 1907) (279 ind.), Onoba kergueleni (E.A. Smith, 1875) (241 ind.), Microdiscula vanhoeffeni Thiele, 1912 123 866 Polar Biol (2014) 37:859–877 Fig. 3 Individual-based interpolation (rarefaction) and extrapolation curves for all the stations sampled at the four areas considered: Cape Adare (*71S, stations A1, A2, A3, A5), Cape Hallett (*72S, stations Hin2, Hin3, Hin4, Hout1, Hout2, Hout4), Coulman Island (*73S, stations C1 and C2) and Terra Nova Bay (*75S, stations R2, R3, R4, SMN). Shaded (or coloured in the on-line version) areas represent 95 % unconditional confidence intervals under the multinomial model Fig. 4 Sample-based interpolation (rarefaction) and extrapolation for incidence data from reference samples corresponding to Cape Adare (*71S, stations A1, A2, A3 and A5 combined), Cape Hallett (*72S, stations Hin2, Hin3, Hin4, Hout1, Hout2 and Hout4 combined) and Terra Nova Bay (*75S, stations R2, R3, R4, SMN). Coulman Island (*73S) stations (i.e. C1 and C2) were not included in this analysis. Shaded (or coloured in the on-line version) areas represent 95 % unconditional confidence intervals under the Bernoulli product model (218 ind.), Adacnarca nitens Pelseneer, 1903 (137 ind.), Lissotesta minutissima (E.A. Smith, 1907) (108 ind.), Solenogastra sp. 4 (104 ind.), Melanella antarctica 123 (Strebel, 1908) (102 ind.) and Lissarca notorcadensis Melvill & Standen, 1907 (101 ind.). The abundances of individuals, standardized to 1,000 m2 hauls (Fig. 7), show a dominance of infaunal groups such as Bivalvia and Scaphopoda in muddy areas (i.e. station C2; Table 1). Epifaunal groups such as Solenogastres, Monoplacophora and Polyplacophora dominate in station Hout4 (sandy bottom; Table 1), while Gastropoda dominate at station A4 (sand and pebbles; Table 1). The multivariate analysis of the assemblage structures (MDS plots) show an apparent separation of northernmost stations (Cape Adare and Cape Hallett, respectively, white triangles and squares in Figs. 8, 9) from southernmost ones (Coulman Island and Cape Russell, respectively, black triangles and squares in Figs. 8, 9) either for the plot referred to ‘‘latitude’’ with all classes combined (squareroot transformed data) (Fig. 8) or for plot where only Bivalvia and Gastropoda have been taken into account (Fig. 9). However, ANOSIM analyses show the existence of a very slight significant difference (at 0.1 %) for the factor ‘‘latitude’’ when square-root transformed data for all classes are considered (Table 5). No statistically significant distinctions can be found for the factor ‘‘depth’’ or for presence/absence data for both factors (Table 5) (MDS plots not reported). Polar Biol (2014) 37:859–877 867 Fig. 5 Ternary plots (Podani and Schmera 2011) showing the additive components of gamma diversity for Cape Adare, Cape Hallett, Terra Nova Bay and the whole data set. Data from Coulman Island (2 stations only) are only reported in Table 4. Smaller ternary plots at the centre of the figure report the single components evaluated (species shared, beta-diversity or species turnover, richness difference, richness agreement, species replacement, nestedness) of the gamma diversity Table 4 Percentage statistics from SDR simplex analysis Additive Cape Adare Cape Hallett Coulman Island Cape Russell Total area Similarity (S) 23.0196 22.4421 16.9811 29.3501 18.9140 Species replacement (R) 31.2392 41.4365 11.3208 45.9922 47.0673 Richness difference (D) 45.7412 36.1214 71.6981 24.6577 34.0187 Beta-diversity (R ? D) 76.9804 77.5579 83.0189 70.6499 81.0860 Richness agreement (R ? S) 54.2588 63.8786 28.3019 75.3423 65.9813 Taxonomical remarks and new records Although further work will be necessary to refine the classifications of about a 18 % of species (which could be either new species or known ones but not easily classifiable due to unavailability of proper iconography in the literature), the remaining 82 % was represented by: (1) new records for the Ross Sea (*12 % of the total number of species) (Figs. 10, 11; Table 6); (2) new species (*6 % of the total); and (3) species already reported in the literature for the Ross Sea (*64 % of the total). Many of the new records are remarkable for a variety of reasons. For example, Tjaernoeia michaeli Engl, 2002 (Fig. 9d) represents not only the first record for the Ross Sea but also the second one after the species description (Engl, 2002). Another species, the monoplacophoran Micropilina arntzi Warén and Hain, 1992 (Fig. 10f), was collected in a total of 38 specimens, 35 of which in a single station (Hout4) showing an ‘‘aggregated distribution’’ as it was already previously observed in the Weddell Sea (Warén and Hain 1992). The family Newtoniellidae, which was not treated by Dell (1990), is present in our samples with two new records for the Ross Sea: Cerithiella seymouriana (Strebel, 1908) (Fig. 9e) and Eumetula dilecta Thiele, 1912 (Fig. 9f). E. dilecta seems to be a rather variable species, and some of our specimens were only tentatively referred to this taxon and reported as E. cf. dilecta (Fig. 9g). Other notable new records for the Ross Sea are Trilirata sexcarinata Warén & Hain, 1996 (Fig. 9l), Pleurotomella deliciosa (Thiele, 1912) (Fig. 9j, k), Onoba egorovae Numanami, 1996 (Fig. 9h) and Onoba paucilirata (Melvill and Standen, 1912) (Fig. 9i). Omalogyra burdwoodiana (Strebel, 1908) (Fig. 9m) could be considered as an ‘‘expected’’ new record, since Omalogyridae, at all 123 868 Polar Biol (2014) 37:859–877 Fig. 6 Relative abundance (number of species) of the six mollusc classes in each sampling station latitudes, can be obtained only when the smallest sediment fraction is examined. Lissotesta similis (Thiele, 1912) is here reported as a new record since our specimens (Fig. 9c) perfectly match the ‘‘syntype 7’’ of this species figured by Engl (2012, plate 26, Fig. 8a). Dell (1990) reports another Lissotesta for the Ross Sea, i.e. L. liratula (Pelseneer, 1903) (Dell 1990, Figs 165–166). These two species, indeed, are very similar and might represent the same taxon, but Dell’s L. liratula has a morphology and sculpture that do not completely fit our one. Engl (2012) considers the original description and the drawing of the holotype of Lissotesta liratula too poor to decide if the two taxa are conspecific or not. Given this state of uncertainty, we have preferred to use the taxon L. similis due to the more similar appearance of our material with the type material, rather than to refer it to the form figured by Dell (1990). The minute species belonging to Brookula were unexpectedly found just with 13 specimens, which were assigned to 3 different species: B. strebeli (Powell, 1951), B. pfefferi (Powell, 1951) (Fig. 9a) and B. cf. argentina Zelaya, Absalão and Pimienta, 2006 (Fig. 9b). Of these species, only B. strebeli (Powell, 1951) was previously reported for the Ross Sea by Dell (1990) who cited it as B. antarctica (now a synonym of B. strebeli). The classification of the species belonging to the Diaphanidae genus Toledonia Dall, 1902 has been problematic since nobody dealt with this group after Dell (1990), even in the case of monographic contributions dedicated to Antarctic molluscs (e.g. Hain 1990; Aldea and Troncoso 2010). In our samples, besides six Toledonia species already known for the Ross Sea, two species were 123 determined at the generic level only. One can be confidently considered new and two others represent new records for the area. These latter two are Toledonia palmeri Dell, 1990, which is known only from the Antarctic Peninsula in shallow water (Fig. 10a), and T. cf. perplexa Dall, 1902 (Fig. 10b), which is reported for the Magellanic area only. T. palmeri has rather distinctive shell forms which, at least on a morphological base, makes the assignment of our specimen to this taxon straightforward. Two other opisthobranchs represent important findings for the Ross Sea. The first one, Philine alata Thiele, 1912 (Fig. 10d, e), although being a rather common species at King George Island and along the Scotia Arc (Engl 2012), has never been previously collected in the Ross Sea. It is possible that this species was confounded with the similar P. apertissima E.A. Smith, 1902, which abounds in the Ross Sea (Schiaparelli, unpublished data). The second one, Prodoridunculus gaussianus Thiele, 1912 (Fig. 10c), is a taxon that will certainly require more study, but the affiliation herein proposed is based on the presence of two rather unique characters: (1) the presence of elevated tubercles on the notum and (2) a series of ‘‘indentations’’ on the notum edge, apparently due to the presence of spicules. Thiele (1912), in the unique figure provided with the description of this species, figures (Tafel XIX, Fig. 5) show a similar feature in the edge of the notum. Within Bivalvia, one Propeamussiidae, Cyclochlamys gaussianus (Thiele, 1912) (Fig. 10g, h) and one Cyclochlamydidae, Cyclochlamys pteriola (Melvill and Standen, 1907) (Fig. 10i, j) can be considered new records for the Ross Sea. C. gaussianus has a distinctive prodissoconch with coarse longitudinal ribs (Fig. 10h), a feature already documented in the literature Polar Biol (2014) 37:859–877 869 diameter of the prodissoconch, which allowed the exclusion of other known forms. This is the case of Limatula ovalis (Fig. 10k–n), which does have a prodissoconch of about 320 lm, while the other similar species, L. hodgsoni and L. simillima have smaller prodissoconchs (Hain and Arnaud 1992). Discussion Fig. 7 Abundance of individuals (standardized to 1,000 m2 hauls) (principal Y axis) and number of species (secondary Y axis) (Aldea and Troncoso 2010, Fig. 221). C. pteriola, although similar to the previous species, has no spiny processes and a completely different microsculpture of prodissoconch and teloeoconch (Fig. 10j) formed by fine radial lyrae. Some of the new records were classified taking into account specific ‘‘indirect’’ morphological features, e.g. the In recent years, the Scientific Committee on Antarctic Research (SCAR) has addressed the necessity for baseline information and long-term monitoring of the Antarctic environment (Turner et al. 2009), while the Committee for the Conservation of Antarctic Marine Living Resources (CCAMLR) has started a bioregionalization program for the Southern Ocean (Grant et al. 2006). In both cases, precise distributional data of species are required in order to develop meaningful models and consistent comparisons across different spatial and temporal scales. In this direction, several projects, above all the Census of Antarctic Marine Life (CAML) (Schiaparelli et al. 2012), have helped in achieving such invaluable information by assessing what is already known and what are the gaps that still exist. Most of this information is now also freely and permanently available through a coordinated network of databases such as the Antarctic Biodiversity Information Facility (ANTABIF, www.antabif.aq). Such an increase in knowledge of Antarctic marine life is unprecedented in the history of Antarctic research (Kaiser et al. 2013) and the new SCAR new program ‘‘AntEco’’ (State of the Antarctic Ecosystem) will have, amongst other tasks, that of analysing spatial patterns of both terrestrial and marine Antarctic species and develop bio-physical models based on the data obtained so far. One of the principal variables used in these studies is species richness, which is an elusive quantity to measure, its perception being confounded by a variety of factors as well as by the scale, extent and grain size examined in a study (Rahbek 2005). Moreover, the comparison of species richness between sites is also a not straightforward task, as biased results can be produced if specific methods, such as rarefaction, are not taken into account (Gotelli and Colwell 2001; Gotelli and Ellison 2013). In Antarctica, amongst the favourite groups of invertebrates used to explore spatial patterns in diversity and richness are the Mollusca (e.g. Clarke et al. 2007). The use of this Phylum as a model taxon is partly due to the large amount of distributional information available compared with other taxa, which also made necessary the creation of a dedicated database for Antarctic species, the Southern Ocean Mollusc Database (SOMBASE, Griffiths et al. 2003). The importance of molluscs relies in the fact that, 123 870 Polar Biol (2014) 37:859–877 Fig. 8 MDS plot for factor latitude of the complete dataset (all classes included) Fig. 9 MDS plot for factor latitude, for Bivalvia and Gastropoda only Table 5 ANOSIM analysis for factors depth and latitude; in bold the unique statistically significative p All classes combined Global R Only Gastropoda and Bivalvia p (%) Global R p (%) Depth (H-transformed data) 0.18 9.4 0.232 5.6 Latitude (H-transformed data) 0.447 0.1 0.278 1.5 Depth (presence–absence data) 0.175 12.6 0.202 6.8 Latitude (presence–absence data) 0.29 1.3 0.2111 3.6 unlike polychaetes or arthropods, whose bodies soon decay after death, a mollusc’s empty shell is informative of the presence of a species in an area, even though this was not collected alive. Mollusc tanatocoenoses (i.e. dead assemblages) can sometimes mirror the original biocoenosis due to their high fidelity (e.g. Carthew and Bosence 1986), but in biodiversity studies, these two fractions, i.e. the living and the dead one, have to be dealt with separately. 123 Unfortunately, in the historical literature where the main focus was species inventory, this distinction was only seldom made explicit and distributional data of both living and dead mollusc accumulated producing the large amount of information we have available today. Regrettably, as a consequence, at least a fraction of the historical distributional data used in biogeographical studies is based on records of dead shells (Schiaparelli, unpublished). Polar Biol (2014) 37:859–877 871 Fig. 10 New records of micromolluscs (Gastropoda) collected by the ‘‘Rauschert dredge’’ in the Ross Sea during the Latitudinal Gradient Program (LGP) voyage on board the RV ‘‘Italica’’. a Brookula pfefferi (scale bar 200 lm), b Brookula cf. argentina (scale bar 500 lm), c Lissotesta similis (scale bar 200 lm), d Tjaernoeia micaeli (scale bar 200 lm), e Cerithiella seymouriana (scale bar 1 mm), f Eumetula dilecta (scale bar 1 mm), g Eumetula cf. dilecta (scale bar 1 mm), h Onoba egorovae (scale bar 500 lm), i Onoba paucilirata (scale bar 500 lm), j Pleurotomella deliciosa (scale bar 500 lm), k Pleurotomella deliciosa protoconch detail relative to j figure (scale bar 200 lm), l Trilirata sexcarinata (scale bar 500 lm), m Omalogyra burdwoodiana (scale bar 200 lm) Despite the considerable amount of data about Antarctic molluscs and notwithstanding the ‘‘contamination’’ of distributional data with records of dead shells, rarefaction curves of species richness plotted against number of sampling locations, still do not show any sign of levelling-off both for Gastropoda and Bivalvia, suggesting that a complete the inventory of molluscs from Antarctic waters is yet to be accomplished (Clarke et al. 2007, Fig. 8). Given this, it would be not surprising to find ‘‘unexpected’’ new records, not only in the historically most neglected areas, where constraints linked to perennial ice cover represent a severe logistic limiting factor for sampling, but also in the more well known ones such as the Ross Sea shelf. For this area, Dell (1990) reported 144 species of Gastropoda (increased to 150 in Griffiths et al. 2009), 42 species of Bivalvia, 3 species of Polyplacophora and 4 species of Scaphopoda. Schiaparelli et al. (2006) studying the samples obtained in the framework of two 2004 research voyages done in the Ross Sea under the LGP framework, where ‘‘standard’’ sampling gears (i.e. grabs, dredges, trawls) were deployed, found 99 species of Gastropoda, 37 species of Bivalvia, 4 species of Polyplacophora and 2 species of Scaphopoda, with a *20 % of species that did not overlapped with the list by Dell (1990) and were considered new records. In the new data set studied here, always collected during the 2004 RV ‘‘Italica’’ expedition but with a ‘‘Rauschert dredge’’ (which was not included in Schiaparelli et al. 123 872 Polar Biol (2014) 37:859–877 Fig. 11 New records of micromolluscs (Gastropoda Opisthobranchia, Bivalvia and Monoplacophora) collected by ‘‘Rauschert dredge’’ in the Ross Sea during the Latitudinal Gradient Program (LGP) voyage on board the RV ‘‘Italica’’. a Toledonia palmeri (scale bar 500 lm), b Toledonia cf. perplexa (scale bar 500 lm), c Prodoridunculus gaussianus (scale bar 200 lm), d Philine alata ventral view (scale bar 1 mm), e Philine alata dorsal view (scale bar 1 mm), f Micropilina arntzi (scale bar 200 lm), g Cyclochlamys gaussianus (scale bar 500 lm), h Cyclochlamys gaussianus protoconch detail relative to g figure (scale bar 100 lm), i Cyclochlamys pteriola (scale bar 1 mm), j Cyclochlamys pteriola protoconch detail relative to i figure (scale bar 100 lm), k Limatula ovalis dorsal view (scale bar 1 mm), l Limatula ovalis ventral view (scale bar 1 mm), m Limatula ovalis protoconch detail relative to k, l figures (scale bar 200 lm), n Limatula ovalis (scale bar 200 lm) 2006), other new records were found and have to be added to the Ross Sea list of species. These account for another 18 % of new findings divided into a 6 % of new species and a 12 % of new records. This means that by considering the present data plus Schiaparelli et al. (2006) and Griffiths et al. (2009), even though we adopt a ‘‘conservative’’ approach, i.e. by excluding the species with an ‘‘uncertain’’ status (*18 % of 123 Polar Biol (2014) 37:859–877 873 Table 6 New records for the Ross Sea and their occurrences at the studied stations A2 A3 A4 Hout1 Hout2 Hout4 Hin3 Hin5 • R3 R4 • • • Cyclochlamys pteriola (Melvill & Standen, 1907) • • • • • Eumetula dilecta Thiele, 1912 • • • • • Lissotesta similis (Pelseneer, 1903) • • Micropilina arntzi Warén & Hain, 1992 • • Omalogyra burdwoodiana Strebel, 1908 • Onoba egorovae Numanami, 1996 Onoba paucilirata (Melvill and Standen, 1912) • • • • Philine alata Thiele, 1912 • • Pleurotomella deliciosa Thiele, 1912 • Prodoridunculus gaussianus Thiele, 1912 • Tjaernoeia micaeli Engl, 2002 • Toledonia cf. perplexa Dall, 1902 Toledonia palmeri Dell, 1990 SMN • Cerithiella similis (Strebel,1908) Limatula ovalis (Thiele, 1912) C2 • Brookula pfefferi Powell, 1951 Brookula cf. argentina Zelaya, Absalao & Pimenta, 2006 Cyclochlamys gaussiana (Thiele, 1912) C1 • Trilirata sexcarinata Warén & Hain, 1996 the total, see results), the overall number of new records of molluscs for the Ross Sea collected during the two LGP voyages undertaken in 2004 corresponds to 51 species of gastropods, 5 species of bivalves and one monoplacophoran. All these new findings (those reported in Schiaparelli et al. 2006 and present data) were achieved thanks to specific sampling strategies: (1) the deployment of a variety of different gears (i.e. grab, sledge and trawl net) in the same area during the RV ‘‘Tangaroa’’ 2004 voyage (Schiaparelli et al. 2006), a strategy that maximized the opportunities of collecting species with a different catchability and (2) the use of a ‘‘Rauchert dredge’’ with a small mesh size during the RV ‘‘Italica’’ 2004 voyage (present data), which allowed to obtain samples of the size that was under-represented in historical samples (Dell 1990). The ‘‘Rauschert dredge’’, analogously to all other towed gears, is not a quantitative method such as grabs or cores. This clearly limits the statistics that can be done, as there is no way to obtain a ‘‘true’’ number of individuals, due to sampling effect. Although it is possible to standardize, i.e. reporting abundances to 1,000 m2 hauls (Rehm et al. 2007), this allows only limited comparisons, and other methods such as rarefaction have to be used. On the other hand, it is evident that the new records and the great amount of specimens found in the ‘‘Rauschert dredge’’ samples are a consequence of the design of the gear which enables a high sampling performance and allows to retain the smallest size fraction of the macrofauna • • • • which was only seldomly studied before in the Ross Sea (Dell 1990). It is well known that the use of a different mesh size, i.e. a smaller one, can strongly affect the outcomes of a study (e.g. Bachelet 1990; Tanaka and Leite 1998; Gage et al. 2002) and that several diversity indices are sensitive both to rare species and to fluctuations in the numbers of specimens per species that characterize each size fraction of a sample (Buckland et al. 2011). For example, Shannon’s H0 generally increases with the decrease of the mesh size until the threshold of 0.5 mm, then it gives a variable response at smaller mesh sizes (Gage et al. 2002). Besides this fact, the variety of sampling gears deployed and the different protocols adopted even between the two 2004 LGP voyages make the comparisons of diversity indices even more complicated. In our samples, H0 values between northernmost and southernmost areas were not found to be statistically different (Fig. 2). Instead, the Simpson index, a more robust diversity measure (Buckland et al. 2011), was found to not to differ between Cape Adare and Cape Hallett, which, however, differed from Coulman Island and Terra Nova Bay (Fig. 2). Indeed, no latitudinal trends in diversity are evident in our data: Terra Nova Bay (the southernmost station) showing higher values of the Simpson index, Cape Adare and Cape Hallett (the northernmost sites) intermediate ones and Coulman Island (which is almost in the middle of the Ross Sea) the lowest values. 123 874 Although species richness is constrained by sampling effort, regardless the considered spatial scale of observation (Clarke et al. 2007; Clarke 2008), sound comparisons can be made using rarefaction methods (Gotelli and Ellison 2013). In our samples, species richness, evaluated through this method, analogously to diversity, was found not to statistically differ in the three main areas considered (Fig. 4, Coulman Island excluded, see ‘‘Materials and methods’’). Here again, although the differences are not statistically relevant, no linear latitudinal trends are evident indeed: the northernmost site (Cape Adare) showed intermediate richness values, the southernmost one (Terra Nova Bay) the lowest and the intermediate site (Cape Hallett) the highest. The high values of beta-diversity (81.1 %) that characterize the whole data set (Table 4; Fig. 5, right lower corner) indicate a high rate of substitution of species between large areas as well as between replicates, suggesting the presence of a high degree of patchiness amongst habitats along the Victoria Land coast. Some of the species also showed a very circumscribed distribution (such as the case of the rare monoplacophoran Micropilina arntzi) and the high number of singletons (21.74 %) and uniques (43.48 %) confirms the elevated habitat heterogeneity. Present data are in apparent contrast from what was observed by Schiaparelli et al. (2006), where a cline in diversity was found, with a higher diversity in the Northern Ross Sea and a lower one in the Southern Ross Sea, and by Clarke et al. (2007, cfr. Fig. 4), where an apparent cline of richness residuals versus plotted latitude was observed. In Schiaparelli et al. (2006), however, much coarser latitudinal groups and ‘‘larger’’ macrofaunal species were taken into account: there, for example, the ‘‘north Ross Sea’’ corresponded to data from Cape Adare (71S) and Cape Hallett (72S) pooled together, while the latitudinal bins here considered have a finer resolution and the latitudes 71S and 72S were analysed separately. Concordantly, the multivariate analyses performed did not resulted in any robust latitudinal patterns. In fact, only the data combined for all classes together after a square root transformation show a slightly significant value of R (at p \ 0.1 %) for the factor latitude. However, these findings have to be interpreted with caution, as it is known that by varying the grain size and extent of an analysis, contrasting results may be obtained (e.g. Rahbek 2005). It should also be considered that all the analyses here performed are based on a total of 18 sampling stations only and that, for some latitudes, only very few samples were available: 5 stations at 71S, 7 at 72S, 2 only at 73S and 4 at 74S (Table 1). This paucity of samples per latitudinal bin did not allow for meaningful statistical tests: rarefaction analyses were limited by the 123 Polar Biol (2014) 37:859–877 availability of only two sampling stations at Coulman Island and multivariate analyses by the low number of possible permutations. It is likely that only the analysis of a much wider data set corrected for the sampling effort (e.g. as in Clarke et al. 2007) where only presence–absence data are taken into account for all the known Ross Sea sampling stations with molluscs (*700 collecting sites; Schiaparelli unpublished) could possibly allow us to understand if a real latitudinal cline or pattern in diversity exists along Victoria Land (Schiaparelli et al. in preparation). Besides the uncertainty that may derive from the low number of samples available in this study, our results are nevertheless in agreement with those of Cummings et al. (2010) who found a similar, nonlinear pattern in diversity along the Victoria Land coast by studying cores obtained from grab samples collected during the RV ‘‘Italica’’ 2004 voyage. Here, a significant correlation between both number of individuals and taxa with sediment phaeophytin concentration across locations indicated a strong linkage between benthos and seafloor productivity (Cummings et al. 2010). In this study, neither latitude nor depth was good predictors of macrofaunal community composition and, overall, there were no indications of simple patterns in Ross Sea benthic assemblage composition (Cummings et al. 2010). Although the grab stations studied by Cummings et al. (2010) and the Rauschert ones studied here are geographically close to each other, it is not possible to use Cummings et al. (2010) phaeophytin and sediment data for our samples due to the fact that a towed gear, when deployed, may collect samples from multiple different communities (e.g. having a different organic content). Consequently, there is no way to find a direct correspondence between species/specimens abundances and other variables, as it is possible with a core from a grab sample. In spite of this, for some areas such as Cape Adare, both gastropod species richness and abundance of individuals in our samples show a striking proportionality to the phaeophytin content in the sediments as reported in Cummings et al. (2010) (compare Fig. 7 this study with Fig. 2b of Cummings et al. 2010). At this site, the highest numbers of molluscs are gastropods such as Onoba turqueti, Powellisetia deserta, Eatoniella kerguelensis, Anatoma euglypta, Microdiscula vanhoeffeni and Onoba kegueleni, all microdetritivores that could be potentially considered indicators of high organic content and, due to the above correspondence, of phaeophytin. Besides the above limitations in statistical analyses, it is clear that a survey done with a ‘‘Rauschert dredge’’ gives access to a size fraction which is species rich, but usually neglected in standard surveys (where larger mesh sizes are used) and is rewarding in terms of new findings. No other gear using the same mesh size, i.e. grabs and corers whose Polar Biol (2014) 37:859–877 samples are generally sieved on small mesh sizes, would have enabled a similar collection of species and specimens due to the very limited area of the bottom sampled at each deployment. All these new findings in the small fraction mirror what are already known for the tropics, where about 33.5 % of the molluscs species have an adult size comprised between 0.4 and 4.1 mm and macromolluscs (i.e. those species which are larger than 41 mm) do account for just 8 % of the total fauna (Bouchet et al. 2002). If we adopt Bouchet’s criterion of size division for our samples (Bouchet et al. 2002), 110 species out of 161 (68.5 %) have a dimension lower or equal to 4.1 mm. In order to obtain similar results in terms of number of sampled specimens and species, it would have been necessary to deploy a grab for an unfeasibly large number of stations/replicates, due to the low densities that shelf molluscs generally have in grab samples (Schiaparelli, unpublished). However, the time demand for such an effort would have been unattainable under standard sampling conditions, especially operating at high latitudes with the well-known logistic constraints. A similar success in sampling with the ‘‘Rauschert dredge’’ during the 2004 expedition is mirrored by the finding of several new species and new distributional records also in other groups. For example, in Isopoda, the 56 % of the recorded species are new to science and 75 species represent new records (Choudhury and Brandt 2009). Unfortunately, few other data sets about molluscs collected by a ‘‘Rauschert dredge’’ are available in the literature, e.g. for the Scotia Arc (Polarstern voyage ANT XIX/ 5: 16 ‘‘Rauschert Dredge’’ stations; Linse et al. 2003) and Bouvet Island (Polarstern voyage ANT XXI/2: with 4 ‘‘Rauschert Dredge’’ stations; Linse 2006). However, direct comparisons of numbers of new records or ratios between different mollusc classes are prevented for different reasons. In Linse et al. (2003), the taxonomic resolution is still at a coarse level since the source is a post voyage data report, and in Linse (2006), a mesh size of 1.5 mm was used, which is three times larger than the one used in the Ross Sea in the present study. Mesh sizes of 500 lm have been instead widely used to sieve grab samples (e.g. Arnaud et al. 2001; Troncoso et al. 2007), but the different sampling methods adopted (qualitative vs quantitative) prevents a meaningful comparison of sampling performance between the two studies. In the present study, the greatest bottleneck was the great amount of laboratory time and ‘‘scientist-hours’’ necessary to sort and classify the studied 8,359 living specimens of molluscs found in the samples. This task took months to be accomplished, often requiring crosschecking of large numbers of specimens in order to achieve consistent division into OTUs or species across samples. This kind of workflow shows similar constraints to those found 875 by researchers working in tropical forest areas (e.g. Lawton et al. 1998). In this phase, straightforward diagnoses were often hampered by the nearly non-existent imagery for these minute species. In order to classify this minute fauna, understand intraspecific variability in shell features, as well as select more robust characters that would allow a ‘‘firstpass’’ discrimination of forms, a large use of SEM images was necessary. This great iconographic effort lead us to document a variety of unknown features such as larval shells, opercula and other morphological features, which will make future classification easier. It is auspicated that, given the high efficiency of sampling of the ‘‘Rauschert dredge’’, more standardized protocols (i.e. use of the same mesh size of 0.5 mm) could be established and applied by different research groups in order to compare, in the future, mollusc assemblages from different Antarctic areas. It is expected that even more species will be added in the near future when samples from the less sampled slope and abyssal areas of the Ross Sea are added. In this regard, it is necessary to point out that in the Ross Sea more than 90 % of the seafloor is deeper than 1,000 m (Griffiths 2010), and most of the stations studied so far rarely exceed 500 m depth. The NIWA ‘‘IPY-CAML’’ voyage (TAN0802) held in the Ross Sea in 2008 during the International Polar Year (2007–2008) sampled deeper water strata in the Ross Sea (i.e. down to 3,490 m) using a ‘‘Brenke-sled’’. These samples, which are currently under study (Schiaparelli et al. unpublished), will allow the refinement of our knowledge of the Ross Sea mollusc fauna and will probably reduce the sampling bias due to the limited number of deep water stations available so far. Acknowledgments The authors are grateful to crew of RV ‘‘Italica’’ for their help and efficiency during the 19th expedition in the Ross Sea, Antarctica. We are indebted to Peter Rehm for the sampling of the ‘‘Rauschert dredge’’ molluscs. We thank the NIWA (National Institute of Water and Atmospheric Research) and the Italian National Antarctic Program (PNRA) for funding and logistic support. We are indebted to Anne-Nina Lörz (NIWA) and two anonymous referees for their comments and suggestions, which greatly improved the MS. This is CAML contribution #162, PNRA project BAMBi (2010/ A1.10) contribution#5 and part of the integrated output from the SCAR-AntEco Science Programme. References Aldea C, Troncoso JS (2010) Moluscos del Mar de Bellingshausen (Antarctica). Troncoso SJ, Aldea C, Vigo Arnaud PM, Troncoso JS, Ramos A (2001) Species diversity and assemblage of macrobenthic Mollusca from the South Shetland Islands and Bransfield Strait (Antarctica). Pol Biol 24:105–112. doi:10.1007/s003000000183 123 876 Bachelet G (1990) The choice of a sieving mesh size in the quantitative assessment of marine macrobenthos: a necessary compromise between aims and constraints. Mar Environ Res 30(1):21–35 Berkman PA, Cattaneo-Vietti R, Chiantore M, Howard-Williams C, Cummings V, Kvitek R (2005) Marine research in the latitudinal gradient project along Victoria Land, Antarctica. Scientia Marina 69:57–63. doi:10.3989/scimar.2005.69s257 Bouchet P, Lozouet P, Maetrati P, Heros V (2002) Assessing the magnitude of species richness in tropical marine environments: exceptionally high numbers of molluscs at a New Caledonia site. Biol J Linnean Soc 75:421–436. doi:10.1046/j.1095-8312.2002. 00052.x Bouchet P, Kantor YI, Sysoev A, Puillandre N (2011) A new operational classification of the Conoidea (Gastropoda). J Molluscan Stud 77:273–308. doi:10.1093/mollus/eyr017 Brenke N (2005) An epibenthic sledge for operations on marine soft bottom and bedrock. Mar Tech Sci J 39:10–19 Buckland ST, Studeny AC, Magurran AE, Newson SE (2011) Biodiversity monitoring: the relevance of detectability. Biological diversity: frontiers in measurement and assessment. Oxford University Press, New York Carthew R, Bosence D (1986) Community preservation in recent shell-gravels, English Channel. Palaeontology 29:243–268 Choudhury M, Brandt A (2007) Composition and distribution of benthic isopod (Crustacea, Malacostraca) families off the Victoria-Land Coast (Ross Sea, Antarctica). Polar Biol 30:1431–1437. doi:10.1007/s00300-007-0304-0 Choudhury M, Brandt A (2009) Benthic isopods (Crustacea, Malacrostaca) from the Ross Sea, Antarctica: species checklist and their zoogeography in the Southern Ocean. Polar Biol 32:599–610. doi:10.1007/s00300-008-0560-7 Choudhury M, Doti BL, Brandt A (2011) Coulmannia rossensis sp.n. (Isopoda, Asellota, Paramunnidae) from the Ross Sea, Southern Ocean. Zookeys 82:45–57. doi:10.3897/zookeys.82.775 Clarke A (2008) Antarctic marine benthic diversity: patterns and processes. J Exp Biol Ecol 366:48–55. doi:10.1016/j.jembe. 2008.07.008 Clarke KR, Gorley RN (2005) PRIMER: getting started with v6. PRIMER-E Ltd, Plymouth Clarke A, Griffiths JH, Linse K, Barnes DKA, Crame AJ (2007) How well do we know the Antarctic marine fauna? A preliminary study of macroecological and biogeographical patterns in Southern Ocean gastropod and bivalve molluscs. Div and Dist 13(5):620–632. doi:10.1111/j.1472-4642.2007.00380.x Colwell RK, Chao A, Gotelli NJ, Lin S-Y, Mao CX, Chazon RL, Longino JT (2012) Models and estimators linking individualbased and sample-based rarefaction, extrapolation and comparison of assemblages. J Plant Ecol 5(1):3–21. doi:10.1093/jpe/ rtr044 Cummings VJ, Thrush SF, Chiantore M, Hewitt JE, Cattaneo-Vietti R (2010) Macrobenthic communities of the north-western Ross Sea shelf: links to depth, sediment characteristics and latitude. Ant Sci 22:793–804. doi:10.1017/S0954102010000489 Dall WH (1902) Illustrations and description of new, unfigured, or imperfectly known shells, chiefly American, in the U.S. National Museum. In: Proceedings of the U.S. National Museum of Natural History 24:499–566 Dayton PK, Oliver JS (1977) Antarctic soft-bottom Benthos in oligotrophic and eutrophic environments. Science 197:55–58. doi:10.1126/science.197.4298.55 Dayton PK, Robilliard GA, Paine RT (1970) Benthos faunal zonation as a result of anchor ice at McMurdo Sound, Antarctica. In: Holgate MW (ed) Antarctic ecology I. Academic, Dordrecht, pp 244–258 123 Polar Biol (2014) 37:859–877 De Broyer C, Danis B (2011) How many species in the Southern Ocean? Towards a dynamic inventory of the Antarctic marine species. Deep-Sea Res II 58:5–17. doi:10.1016/j.dsr2.2010.10.007 Dell RK (1990) Antarctic Mollusca. With special reference to the fauna of the Ross Sea. J R Soc N Z Bull 27:1–311 Engl W (2002) Tjaernoeia michaeli sp. nov. La Conchiglia 34(303): 11–12 Engl W (2012) Shalls of Antarctica. ConchBooks, Germany Gage JD, Hughes DJ, Gonzalez Vecino JL (2002) Sieve size influence in estimating biomass, abundance and diversity in samples of deep-sea macrobenthos. Mar Ecol Progr Ser 225:97–107. doi:10. 3354/meps225097 Ghiglione C, Alvaro MC, Griffiths HJ, Linse K, Schiaparelli S (2013) Ross Sea Mollusca from the latitudinal gradient program: R/V Italica 2004 Rauschert dredge samples. Zookeys 341:37–48. doi:10.3897/zookeys.341.6031 Gotelli NJ, Colwell RK (2001) Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecol Lett 4:379–391 Gotelli NJ, Ellison AM (2013) The measurement of biodiversity. In: Gotelli NJ, Ellison AM (eds) A primer of ecological statistics, 2nd edn. Sinauer Associates, Sunderland, pp 449–482 Grant S, Constable A, Raymond B et al (2006) Bioregionalisation of the Southern Ocean: report of experts workshop, Hobart Griffiths HJ (2010) Antarctic marine biodiversity—what do we know about the distribution of life in the Southern Ocean? PLoS ONE 5(8):e11683. doi:10.1371/journal.pone.0011683 Griffiths HJ, Linse K, Crame AJ (2003) SOMBASE—Southern Ocean Mollusc Database: a tool for biogeographic analysis in diversity and ecology. Org Divers Evol 3:207–213 Griffiths HJ, Barnes DKA, Linse K (2009) Towards a generalized biogeography of the Southern Ocean benthos. J Biogeogr 36:162–177. doi:10.1111/j.1365-2699.2008.01979.x Griffiths HJ, Danis B, Clarke A (2011) Quantifying Antarctic marine biodiversity: the SCAR-MarBIN data portal. Deep-Sea Res II 58:18–29. doi:10.1016/j.dsr2.2010.10.008 Hain S (1990) The benthic seashells (Gastropoda and Bivalvia) of the Weddell Sea, Antarctica. Ber Polarforsch 70:1–181 Hain S, Arnaud P (1992) Notes on the reproduction of high-Antarctic molluscs from the Weddell Sea. Pol Biol 12:303–312 Hsieh TC, Ma KH, Chao A (2013) iNEXT online: interpolation and extrapolation (Version 1.3.0) [Software]. http://chao.stat.nthu. edu.tw/blog/software-download/. Accessed 15 Jan 2014 Kaiser S, Brandão SN, Brix S, Barnes DKA, Bowden DA, Ingels J, Leese F, Schiaparelli S, Arango CP, Badhe R, Bax N, Blazewicz-Paszkowycz M, Brandt A, Brenke N, Catarino AI, David B, De Ridder C, Dubois P, Ellingsen KE, Glover AG, Griffiths HJ, Gutt J, Halanych KM, Havermans C, Held C, Janussen D, Lörz AN, Pearce DA, Pierrat B, Riehl T, Rose A, Sands CJ, Soler-Membrives A, Schüller M, Strugnell JM, Vanreusel A, Veit-Köhler G, Wilson NG, Yasuhara M (2013) Patterns, processes and vulnerability of Southern Ocean benthos: a decadal leap in knowledge and understanding. Mar Biol 160:2295–2317. doi:10.1007/s00227-013-2232-6 Lamy E (1905) Gastéropodes Prosobranches recueillis par l’Expédition Antartique Française du Dr. Charcot. Bull Mus Hist Nat 11:475–483 Lawton JH, Bignell DE, Bolton B, Bloemers GF, Eggleton P, Hammond PM, Hodda M, Holt RD, Larsen TB, Mawdsley NA, Stork NE, Srivastava DS, Watt AD (1998) Biodiversity inventories, indicator taxa and effects of habitat modification in tropical forest. Nature 391:72–76 Linse K (2006) New records of shelled marine molluscs at Bouvet Island and preliminary assessment of their biogeographic affinities. Pol Biol 29:120–127. doi:10.1007/s00300-005-0721-x Polar Biol (2014) 37:859–877 Linse K, Schrödl M, Zelaya D (2003) Biodiversity, biogeography and evolution of Mollusca. Ber Polarforsch 462:19–28 Lörz AN, di Renzo A, Nickel J (1999) Comparative analysis of three sampling gear types for marine macrobenthos. Ber Polarforsch 330:134–151 Lörz AN, Kaiser S, Bowden DA (2013) Macrofaunal crustaceans in the benthic boundary layer from shelf break to abyssal depths in the Ross Sea (Antarctica). Pol Biol 36:445–451. doi:10.1007/ s00300-012-1269-1 Melvill JC, Standen R (1907) The marine Mollusca of the Scottish National Antarctic expedition. Trans R Soc Edinb 46:119–157 Melvill JC, Standen R (1912) The marine Mollusca of the Scottish National Antarctic expedition. Part II. Trans R Soc Edinb 46:333–366 Numanami H (1996) Taxonomic study on Antarctic gastropods collected by Japanese Antarctic research expeditions. Mem Natl Inst Polar Res 39:1–245 Pelseneer P (1903) Mollusques (Amphineures, Gastéropodes et Lamellibranches). Résultat du Voyage du S.Y. ‘‘Belgica’’ en 1897–1898–1899. Rapports Scientifiques, Zoologie Podani J (2001) SYN-TAX 2000. Computer programs for data analysis in ecology and systematics. Scientia Publishing, Budapest Podani J, Schmera D (2011) A new conceptual and methodological framework for exploring and explaining pattern in presence– absence data. Oikos 120:1625–1638. doi:10.1111/j.1600-0706. 2011.19451.x Powell AWB (1951) Antarctic and subantarctic Mollusca: Pelecypoda and Gastropoda. Discov Report 26:47–196 Rahbek C (2005) The role of spatial scale and the perception of largescale species-richness patterns. Ecol Lett 8:224–239. doi:10. 1111/j.1461-0248.2004.00701.x Rehm P, Heard WR (2008) Leucon (Crymoleucon) rossi, a new species (Crustacea: Cumacea: Leuconidae) from the shelf waters of the Ross Sea (Antarctica), with a key to the genus Leucon south of 60S. Sci Mar 72(4):683–691 Rehm P, Thatje S, Arntz EW, Brandt A, Heilmayer O (2006) Distribution and composition of macrozoobenthic communities along a Victoria-Land Transect (Ross Sea, Antarctica). Polar Biol 29:782–790 Rehm P, Thatje S, Brandt A, Siegel MU (2007) Composition and distribution of the peracarid crustacean fauna along a latitudinal transect off Victoria Land (Ross Sea, Antarctica) with special emphasis on the Cumacea. Polar Biol 30:871–881 877 Schiaparelli S, Lorz AN, Cattaneo-Vietti R (2006) Diversity and distribution of mollusc assemblages on the Victoria Land coast and the Balleny Islands, Ross Sea, Antarctica. Ant Sci 18(4):615–631 Schiaparelli S, Danis B, Wadley V, Stoddart DM (2012) The census of antarctic marine life: the first available baseline for antarctic marine biodiversity. In: Verde C, Di Prisco G (eds) Adaptation and evolution in marine environments. Spinger, Berlin, pp 3–20 Schrödl M, Bohn JM, Brenke N, Rolán E, Schwabe E (2011) Abundance, diversity, and latitudinal gradients of southeastern Atlantic and Antarctic abyssal gastropods. Deep Sea Res II 58:49–57 Smith EA (1875) Description of some new shell from Kerguelen’s Island. Ann Mag Nat Hist 16:67–73 Smith EA (1902) Mollusca. Report on the collections of natural history, made in the antarctic region during the voyage of the ‘‘Southern Cross’’. Br Mus (Natl Hist) 201–213 Smith EA (1907) Mollusca, part V Lamellibranchiata. National antarctic expedition 1901–1904. Nat Hist Rep Zool 2:1–7 Strebel H (1908) Die Gastropoden. Wiss. Ergebn. Schwed. Süd-PolExped (1901–1903) 6:1–112 Tanaka MO, Leite PPP (1998) The effect of sieve mesh size on the abundance and composition of macrophyte-associated macrofaunal assemblages. Hydrobiologia 389(1–3):21–28. doi:10.1023/ A:1003589404593 Thiele J (1912) Die Antarktischen Schnecken Und Muschlen. Berlin, Germany Troncoso SJ, Aldea C, Arnaud P, Ramos A, Garcia F (2007) Quantitative analysis of soft-bottom molluscs in the Bellingshausen Sea and around Peter I Island. Polar Res 26:126–134 Turner J, Convey P, di Prisco G et al (2009) Antarctic climate change and the environment. Cambridge Warén A, Hain S (1992) Laevipilina antarctica and Micropilina arntzi, two new monoplacophorans from the Antarctic. Veliger 35:165–176 Warén A, Hain S (1996) Description of Zerotulidae fam. Nov. (Littorinoidea), with comments on Antarctic Littorinid gastropods. Veliger 39:277–334 Zelaya DG, Absalão RS, Pimenta AD (2006) A revision of Benthobrookula Clarke, 1961 (Gastropoda, Trochoidea) in the southwestern Atlantic Ocean. J Molluscan Stud 72:77–87 123 View publication stats